nips nips2004 nips2004-85 knowledge-graph by maker-knowledge-mining

85 nips-2004-Instance-Based Relevance Feedback for Image Retrieval

Source: pdf

Author: Giorgio Gia\-cin\-to, Fabio Roli

Abstract: High retrieval precision in content-based image retrieval can be attained by adopting relevance feedback mechanisms. These mechanisms require that the user judges the quality of the results of the query by marking all the retrieved images as being either relevant or not. Then, the search engine exploits this information to adapt the search to better meet user’s needs. At present, the vast majority of proposed relevance feedback mechanisms are formulated in terms of search model that has to be optimized. Such an optimization involves the modification of some search parameters so that the nearest neighbor of the query vector contains the largest number of relevant images. In this paper, a different approach to relevance feedback is proposed. After the user provides the first feedback, following retrievals are not based on knn search, but on the computation of a relevance score for each image of the database. This score is computed as a function of two distances, namely the distance from the nearest non-relevant image and the distance from the nearest relevant one. Images are then ranked according to this score and the top k images are displayed. Reported results on three image data sets show that the proposed mechanism outperforms other state-of-the-art relevance feedback mechanisms. 1 In t rod u ct i on A large number of content-based image retrieval (CBIR) systems rely on the vector representation of images in a multidimensional feature space representing low-level image characteristics, e.g., color, texture, shape, etc. [1]. Content-based queries are often expressed by visual examples in order to retrieve from the database the images that are “similar” to the examples. This kind of retrieval is often referred to as K nearest-neighbor retrieval. It is easy to see that the effectiveness of content-based image retrieval systems (CBIR) strongly depends on the choice of the set of visual features, on the choice of the “metric” used to model the user’s perception of image similarity, and on the choice of the image used to query the database [1]. Typically, if we allow different users to mark the images retrieved with a given query as relevant or non-relevant, different subsets of images will be marked as relevant. Accordingly, the need for mechanisms to adapt the CBIR system response based on some feedback from the user is widely recognized. It is interesting to note that while relevance feedback mechanisms have been first introduced in the information retrieval field [2], they are receiving more attention in the CBIR field (Huang). The vast majority of relevance feedback techniques proposed in the literature is based on modifying the values of the search parameters as to better represent the concept the user bears in mind. To this end, search parameters are computed as a function of the relevance values assigned by the user to all the images retrieved so far. As an example, relevance feedback is often formulated in terms of the modification of the query vector, and/or in terms of adaptive similarity metrics. [3]-[7]. Recently, pattern classification paradigms such as SVMs have been proposed [8]. Feedback is thus used to model the concept of relevant images and adjust the search consequently. Concept modeling may be difficult on account of the distribution of relevant images in the selected feature space. “Narrow domain” image databases allows extracting good features, so that images bearing similar concepts belong to compact clusters. On the other hand, “broad domain” databases, such as image collection used by graphic professionals, or those made up of images from the Internet, are more difficult to subdivide in cluster because of the high variability of concepts [1]. In these cases, it is worth extracting only low level, non-specialized features, and image retrieval is better formulated in terms of a search problem rather then concept modeling. The present paper aims at offering an original contribution in this direction. Rather then modeling the concept of “relevance” the user bears in mind, feedback is used to assign each image of the database a relevance score. Such a score depends only from two dissimilarities (distances) computed against the images already marked by the user: the dissimilarity from the set of relevant images, and the dissimilarity from the set of non-relevant images. Despite its computational simplicity, this mechanism allows outperforming state-of-the-art relevance feedback mechanisms both on “narrow domain” databases, and on “broad domain” databases. This paper is organized as follows. Section 2 illustrates the idea behind the proposed mechanism and provides the basic assumptions. Section 3 details the proposed relevance feedback mechanism. Results on three image data sets are presented in Section 4, where performances of other relevance feedback mechanisms are compared. Conclusions are drawn in Section 5. 2 In st an ce- b ased rel evan ce est i m at i on The proposed mechanism has been inspired by classification techniques based on the “nearest case” [9]-[10]. Nearest-case theory provided the mechanism to compute the dissimilarity of each image from the sets of relevant and non–relevant images. The ratio between the nearest relevant image and the nearest non-relevant image has been used to compute the degree of relevance of each image of the database [11]. The present section illustrates the rationale behind the use of the nearest-case paradigm. Let us assume that each image of the database has been represented by a number of low-level features, and that a (dis)similarity measure has been defined so that the proximity between pairs of images represents some kind of “conceptual” similarity. In other words, the chosen feature space and similarity metric is meaningful at least for a restricted number of users. A search in image databases is usually performed by retrieving the k most similar images with respect to a given query. The dimension of k is usually small, to avoid displaying a large number of images at a time. Typical values for k are between 10 and 20. However, as the “relevant” images that the user wishes to retrieve may not fit perfectly with the similarity metric designed for the search engine, the user may be interested in exploring other regions of the feature space. To this end, the user marks the subset of “relevant” images out of the k retrieved. Usually, such relevance feedback is used to perform a new k-nn search by modifying some search parameters, i.e., the position of the query point, the similarity metric, and other tuning parameters [1]-[7]. Recent works proposed the use of support vector machine to learn the distribution of relevant images [8]. These techniques require some assumption about the general form of the distribution of relevant images in the feature space. As it is difficult to make any assumption about such a distribution for broad domain databases, we propose to exploit the information about the relevance of the images retrieved so far in a nearest-neighbor fashion. Nearest-neighbor techniques, as used in statistical pattern recognition, case-based reasoning, or instance-based learning, are effective in all applications where it is difficult to produce a high-level generalization of a “class” of objects [9]-[10],[12][13]. Relevance learning in content base image retrieval may well fit into this definition, as it is difficult to provide a general model that can be adapted to represent different concepts of similarity. In addition, the number of available cases may be too small to estimate the optimal set of parameters for such a general model. On the other hand, it can be more effective to use each “relevant” image as well as each “non-relevant” image, as “cases” or “instances” against which the images of the database should be compared. Consequently, we assume that an image is as much as relevant as much as its dissimilarity from the nearest relevant image is small. Analogously, an image is as much as non-relevant as much as its dissimilarity from the nearest non-relevant image is small. 3 Rel evan ce S core Com p u t ati on According to previous section, each image of the database can be thus characterized by a “degree of relevance” and a “degree of non-relevance” according to the dissimilarities from the nearest relevant image, and from the nearest non-relevant image, respectively. However, it should be noted that these degrees should be treated differently because only “relevant” images represent a “concept” in the user’s mind, while “non-relevant” images may represent a number of other concepts different from user’s interest. In other words, while it is meaningful to treat the degree of relevance as a degree of membership to the class of relevant images, the same does not apply to the degree of non-relevance. For this reason, we propose to use the “degree of non-relevance” to weight the “degree of relevance”. Let us denote with R the subset of indexes j ∈ {1,...,k} related to the set of relevant images retrieved so far and the original query (that is relevant by default), and with NR the subset of indexes j ∈ (1,...,k} related to the set of non-relevant images retrieved so far. For each image I of the database, according to the nearest neighbor rule, let us compute the dissimilarity from the nearest image in R and the dissimilarity from the nearest image in NR. Let us denote these dissimilarities as dR(I) and dNR(I), respectively. The value of dR(I) can be clearly used to measure the degree of relevance of image I, assuming that small values of dR(I) are related to very relevant images. On the other hand, the hypothesis that image I is relevant to the user’s query can be supported by a high value of dNR(I). Accordingly, we defined the relevance score ! dR ( I ) $ relevance ( I ) = # 1 + dN ( I ) &

Reference: text

Summary: the most important sentenses genereted by tfidf model

sentIndex sentText sentNum sentScore

1 it Abstract High retrieval precision in content-based image retrieval can be attained by adopting relevance feedback mechanisms. [sent-3, score-1.649]

2 These mechanisms require that the user judges the quality of the results of the query by marking all the retrieved images as being either relevant or not. [sent-4, score-0.948]

3 Then, the search engine exploits this information to adapt the search to better meet user’s needs. [sent-5, score-0.207]

4 At present, the vast majority of proposed relevance feedback mechanisms are formulated in terms of search model that has to be optimized. [sent-6, score-1.018]

5 Such an optimization involves the modification of some search parameters so that the nearest neighbor of the query vector contains the largest number of relevant images. [sent-7, score-0.592]

6 In this paper, a different approach to relevance feedback is proposed. [sent-8, score-0.734]

7 After the user provides the first feedback, following retrievals are not based on knn search, but on the computation of a relevance score for each image of the database. [sent-9, score-0.926]

8 This score is computed as a function of two distances, namely the distance from the nearest non-relevant image and the distance from the nearest relevant one. [sent-10, score-0.69]

9 Images are then ranked according to this score and the top k images are displayed. [sent-11, score-0.292]

10 Reported results on three image data sets show that the proposed mechanism outperforms other state-of-the-art relevance feedback mechanisms. [sent-12, score-1.102]

11 1 In t rod u ct i on A large number of content-based image retrieval (CBIR) systems rely on the vector representation of images in a multidimensional feature space representing low-level image characteristics, e. [sent-13, score-0.91]

12 Content-based queries are often expressed by visual examples in order to retrieve from the database the images that are “similar” to the examples. [sent-17, score-0.376]

13 This kind of retrieval is often referred to as K nearest-neighbor retrieval. [sent-18, score-0.253]

14 Typically, if we allow different users to mark the images retrieved with a given query as relevant or non-relevant, different subsets of images will be marked as relevant. [sent-20, score-0.86]

15 Accordingly, the need for mechanisms to adapt the CBIR system response based on some feedback from the user is widely recognized. [sent-21, score-0.567]

16 It is interesting to note that while relevance feedback mechanisms have been first introduced in the information retrieval field [2], they are receiving more attention in the CBIR field (Huang). [sent-22, score-1.126]

17 The vast majority of relevance feedback techniques proposed in the literature is based on modifying the values of the search parameters as to better represent the concept the user bears in mind. [sent-23, score-1.312]

18 To this end, search parameters are computed as a function of the relevance values assigned by the user to all the images retrieved so far. [sent-24, score-1.014]

19 As an example, relevance feedback is often formulated in terms of the modification of the query vector, and/or in terms of adaptive similarity metrics. [sent-25, score-1.04]

20 Feedback is thus used to model the concept of relevant images and adjust the search consequently. [sent-28, score-0.487]

21 Concept modeling may be difficult on account of the distribution of relevant images in the selected feature space. [sent-29, score-0.483]

22 “Narrow domain” image databases allows extracting good features, so that images bearing similar concepts belong to compact clusters. [sent-30, score-0.602]

23 On the other hand, “broad domain” databases, such as image collection used by graphic professionals, or those made up of images from the Internet, are more difficult to subdivide in cluster because of the high variability of concepts [1]. [sent-31, score-0.544]

24 In these cases, it is worth extracting only low level, non-specialized features, and image retrieval is better formulated in terms of a search problem rather then concept modeling. [sent-32, score-0.627]

25 Rather then modeling the concept of “relevance” the user bears in mind, feedback is used to assign each image of the database a relevance score. [sent-34, score-1.341]

26 Such a score depends only from two dissimilarities (distances) computed against the images already marked by the user: the dissimilarity from the set of relevant images, and the dissimilarity from the set of non-relevant images. [sent-35, score-0.729]

27 Despite its computational simplicity, this mechanism allows outperforming state-of-the-art relevance feedback mechanisms both on “narrow domain” databases, and on “broad domain” databases. [sent-36, score-0.917]

28 Section 2 illustrates the idea behind the proposed mechanism and provides the basic assumptions. [sent-38, score-0.201]

29 Results on three image data sets are presented in Section 4, where performances of other relevance feedback mechanisms are compared. [sent-40, score-1.131]

30 2 In st an ce- b ased rel evan ce est i m at i on The proposed mechanism has been inspired by classification techniques based on the “nearest case” [9]-[10]. [sent-42, score-0.388]

31 Nearest-case theory provided the mechanism to compute the dissimilarity of each image from the sets of relevant and non–relevant images. [sent-43, score-0.577]

32 The ratio between the nearest relevant image and the nearest non-relevant image has been used to compute the degree of relevance of each image of the database [11]. [sent-44, score-1.606]

33 Let us assume that each image of the database has been represented by a number of low-level features, and that a (dis)similarity measure has been defined so that the proximity between pairs of images represents some kind of “conceptual” similarity. [sent-46, score-0.561]

34 A search in image databases is usually performed by retrieving the k most similar images with respect to a given query. [sent-48, score-0.648]

35 The dimension of k is usually small, to avoid displaying a large number of images at a time. [sent-49, score-0.209]

36 However, as the “relevant” images that the user wishes to retrieve may not fit perfectly with the similarity metric designed for the search engine, the user may be interested in exploring other regions of the feature space. [sent-51, score-0.887]

37 To this end, the user marks the subset of “relevant” images out of the k retrieved. [sent-52, score-0.404]

38 Usually, such relevance feedback is used to perform a new k-nn search by modifying some search parameters, i. [sent-53, score-0.917]

39 , the position of the query point, the similarity metric, and other tuning parameters [1]-[7]. [sent-55, score-0.265]

40 Recent works proposed the use of support vector machine to learn the distribution of relevant images [8]. [sent-56, score-0.41]

41 These techniques require some assumption about the general form of the distribution of relevant images in the feature space. [sent-57, score-0.465]

42 As it is difficult to make any assumption about such a distribution for broad domain databases, we propose to exploit the information about the relevance of the images retrieved so far in a nearest-neighbor fashion. [sent-58, score-0.932]

43 Relevance learning in content base image retrieval may well fit into this definition, as it is difficult to provide a general model that can be adapted to represent different concepts of similarity. [sent-60, score-0.615]

44 On the other hand, it can be more effective to use each “relevant” image as well as each “non-relevant” image, as “cases” or “instances” against which the images of the database should be compared. [sent-62, score-0.537]

45 Consequently, we assume that an image is as much as relevant as much as its dissimilarity from the nearest relevant image is small. [sent-63, score-0.933]

46 Analogously, an image is as much as non-relevant as much as its dissimilarity from the nearest non-relevant image is small. [sent-64, score-0.641]

47 However, it should be noted that these degrees should be treated differently because only “relevant” images represent a “concept” in the user’s mind, while “non-relevant” images may represent a number of other concepts different from user’s interest. [sent-66, score-0.488]

48 In other words, while it is meaningful to treat the degree of relevance as a degree of membership to the class of relevant images, the same does not apply to the degree of non-relevance. [sent-67, score-0.72]

49 ,k} related to the set of relevant images retrieved so far and the original query (that is relevant by default), and with NR the subset of indexes j ∈ (1,. [sent-72, score-0.831]

50 ,k} related to the set of non-relevant images retrieved so far. [sent-75, score-0.313]

51 For each image I of the database, according to the nearest neighbor rule, let us compute the dissimilarity from the nearest image in R and the dissimilarity from the nearest image in NR. [sent-76, score-1.22]

52 The value of dR(I) can be clearly used to measure the degree of relevance of image I, assuming that small values of dR(I) are related to very relevant images. [sent-78, score-0.84]

53 On the other hand, the hypothesis that image I is relevant to the user’s query can be supported by a high value of dNR(I). [sent-79, score-0.533]

54 dR ( I ) $ relevance ( I ) = # 1 + dN ( I ) & " % '1 (1) This formulation of the score can be easily explained in terms of a distanceweighted 2-nn estimation of the posterior probability that image I is relevant. [sent-81, score-0.705]

55 The 2 nearest neighbors are made up of the nearest relevant image, and the nearest nonrelevant image, while the weights are computed as the inverse of the distance from the nearest neighbors. [sent-82, score-0.678]

56 The relevance score computed according to equation (1) is then used to rank the images and the first k are presented to the user. [sent-83, score-0.745]

57 4 Exp eri m en t al resu l t s In order to test the proposed method and compare it with other methods described in the literature, three image databases have been used: the MIT database, a database contained in the UCI repository, and a subset of the Corel database. [sent-84, score-0.511]

58 These databases are currently used for assessing and comparing relevance feedback techniques [5],[7],[14]. [sent-85, score-0.914]

59 This database contains 40 texture images that have been manually classified into fifteen classes. [sent-90, score-0.397]

60 Each of these images has been subdivided into sixteen non-overlapping images, obtaining a data set with 640 images. [sent-91, score-0.287]

61 Sixteen Gabor filters were used to characterise these images, so that each image is represented by a 16-dimensional feature vector [14]. [sent-92, score-0.287]

62 The images are subdivided into seven data classes (brickface, sky, foliage, cement, window, path, and grass). [sent-98, score-0.279]

63 The database extracted from the Corel collection is available at the KDD-UCI repository (http://kdd. [sent-101, score-0.203]

64 For the first two databases, each image is used as a query, while for the Corel database, 500 images have been randomly extracted and used as query, so that all the 43 classes are represented. [sent-111, score-0.468]

65 Relevance feedback is performed by marking images belonging to the same class of the query as relevant, and all other images as non-relevant. [sent-113, score-0.954]

66 The user’s query itself is included in the set of relevant images. [sent-114, score-0.338]

67 Results are evaluated in term of the retrieval precision averaged over all the considered queries. [sent-116, score-0.407]

68 The precision is measured as the fraction of relevant images contained in the 20 top retrieved images. [sent-117, score-0.613]

69 As the first two databases are of the “narrow domain” type, while the third is of the “broad domain” type, this experimental set-up allowed a thorough testing of the proposed technique. [sent-118, score-0.209]

70 These two methods have been selected because they can be easily implemented, and their performances can be compared to those provided by a large number of relevance feedback techniques proposed in the CBIR literature (see for example results presented in [15]). [sent-120, score-1.005]

71 1 Experiments w ith th e MI T database This database can be considered of the “narrow domain” type as it contains only images of textures of 40 different types. [sent-124, score-0.53]

72 Figure 1 show the performances of the proposed relevance feedback mechanism and those of the two techniques used for comparison. [sent-126, score-1.096]

73 100 % Precision 95 90 Relevance Score Bayes QS 85 MindReader 80 75 0 rf 1 rf 2 rf 3 rf 4 rf 5 rf Iter. [sent-127, score-1.518]

74 Feedback 6 rf 7 rf 8 rf Figure 1: Retrieval Performances for the MIT database in terms of average percentage retrieval precision. [sent-129, score-1.181]

75 After the first feedback iteration (1rf in the graph), each relevance feedback mechanism is able to improve the average precision attained in the first retrieval by more than 10%, the proposed mechanism performing slightly better than MindReader. [sent-130, score-1.883]

76 However, if the user aims to better refine the search by additional feedback iteration, MindReader and Bayes QS are not able to exploit the additional information, as they provide no improvements after the second feedback iteration. [sent-132, score-0.915]

77 On the other hand, the proposed mechanism provides further improvement in precision by increasing the number of iteration. [sent-133, score-0.327]

78 These improvements are very small because the first feedback already provides a high precision value, near to 95%. [sent-134, score-0.487]

79 2 Experiments w ith th e UC I database This database too can be considered of the “narrow domain” type as the images clearly belong to one of the seven data classes, and features have been extracted accordingly. [sent-136, score-0.646]

80 100 % Precision 98 96 Relevance Score Bayes QS 94 MindReader 92 90 0 rf 1 rf 2 rf 3 rf 4 rf 5 rf 6 rf 7 rf 8 rf Iter. [sent-137, score-2.277]

81 Feedback Figure 2: Retrieval Performances for the UCI data set in terms of average percentage retrieval precision. [sent-139, score-0.289]

82 Nonetheless, each of the considered mechanism is able to exploit relevance feedback, Mindreader and Bayes QS providing a 6% improvement, while the proposed mechanism attains a 8% improvement. [sent-142, score-0.748]

83 This example clearly shows the superiority of the proposed technique, as it attains a precision of 99% after the second iteration. [sent-143, score-0.304]

84 On the other hand, Bayes QS also exploits further feedback iteration attaining a precision of 98% after 7 iterations, while MindReader does not improve the precision attained after the first iteration. [sent-145, score-0.804]

85 As the user typically allows very few feedback iterations, the proposed mechanism proved to be very suited for narrow domain databases as it allows attaining a precision close to 100%. [sent-146, score-1.148]

86 3 Experiments w ith th e Co rel databas e Figures 3 and 4 show the performances attained on two feature sets extracted from the Corel database. [sent-148, score-0.398]

87 This database is of the “broad domain” type as images represent a very large number of concepts, and the selected feature sets represent conceptual similarity between pairs of images only partly. [sent-149, score-0.697]

88 Let us note that the retrieval precision after the first k-nn search (0rf in the graphs) is quite small. [sent-151, score-0.512]

89 This is a consequence of the difficulty of selecting a good feature space to represent conceptual similarity between pairs of images in a broad domain database. [sent-152, score-0.508]

90 This difficulty is partially overcome by using MindReader or Bayes QS as they allow improving the retrieval precision by 10% to 15% according to the number of iteration allowed, and according to the selected feature space. [sent-153, score-0.528]

91 Let us recall that both MindReader and Bayes QS perform a query movement in order to perform a k-nn query on a more promising region of the feature space. [sent-154, score-0.442]

92 On the other hand, the proposed mechanism based on ranking all the images of the database according to a relevance score, not only provided higher precision after the first feedback, but also allow to improve significantly the retrieval precision as the number of iteration is increased. [sent-155, score-1.561]

93 As the initial precision is quite small, a user may have more willingness to perform further iterations as the proposed mechanism allows retrieving new relevant images. [sent-156, score-0.728]

94 Figure 3: Retrieval Performances for the Corel data set (Color Moments feature set) in terms of average percentage retrieval precision Figure 4: Retrieval Performances for the Corel data set (Co-occurrence Texture feature set) in terms of average percentage retrieval precision. [sent-157, score-0.848]

95 5 Con cl u si on s In this paper, we proposed a novel relevance feedback technique for content-based image retrieval. [sent-158, score-0.984]

96 This is the case of relevance feedback in CBIR, where the use of classification models should require a suitable formulation in order to avoid socalled “small sample” problems. [sent-162, score-0.774]

97 Reported results clearly showed the superiority of the proposed mechanism especially when large databases made up of images related to many different concepts are searched. [sent-163, score-0.645]

98 : Content-based image retrieval at the end of the early years. [sent-172, score-0.448]

99 : Relevance feedback in image retrieval: a comprehensive review, Multimedia Systems 8(6) (2003) 536-544 [9] Aha D. [sent-210, score-0.502]

100 , Probabilistic feature relevance learning for content-based image retrieval, Computer Vision and Image Understanding 75 (1999) 150-164. [sent-240, score-0.68]

similar papers computed by tfidf model

tfidf for this paper:

wordName wordTfidf (topN-words)

[('relevance', 0.427), ('feedback', 0.307), ('rf', 0.253), ('retrieval', 0.253), ('images', 0.209), ('image', 0.195), ('user', 0.195), ('query', 0.192), ('mindreader', 0.191), ('qs', 0.169), ('precision', 0.154), ('relevant', 0.146), ('performances', 0.137), ('nearest', 0.133), ('database', 0.133), ('databases', 0.128), ('cbir', 0.127), ('dissimilarity', 0.118), ('mechanism', 0.118), ('retrieved', 0.104), ('corel', 0.094), ('score', 0.083), ('search', 0.079), ('narrow', 0.075), ('dr', 0.074), ('difficult', 0.07), ('concepts', 0.07), ('domain', 0.069), ('mechanisms', 0.065), ('attained', 0.06), ('feature', 0.058), ('dissimilarities', 0.055), ('proposed', 0.055), ('texture', 0.055), ('broad', 0.053), ('concept', 0.053), ('techniques', 0.052), ('rel', 0.05), ('degree', 0.049), ('uci', 0.049), ('similarity', 0.048), ('subdivided', 0.047), ('attaining', 0.047), ('bayes', 0.046), ('cagliari', 0.042), ('dnr', 0.042), ('giacinto', 0.042), ('lew', 0.042), ('modification', 0.042), ('multimedia', 0.042), ('roli', 0.042), ('santini', 0.042), ('huang', 0.042), ('superiority', 0.042), ('metric', 0.042), ('conceptual', 0.04), ('classification', 0.04), ('extracted', 0.038), ('vast', 0.037), ('retrieving', 0.037), ('marking', 0.037), ('evan', 0.037), ('percentage', 0.036), ('ce', 0.036), ('jain', 0.034), ('characterise', 0.034), ('retrieve', 0.034), ('indexes', 0.034), ('features', 0.032), ('repository', 0.032), ('iteration', 0.032), ('rationale', 0.031), ('sixteen', 0.031), ('bears', 0.031), ('difficulty', 0.031), ('th', 0.031), ('attains', 0.03), ('behind', 0.028), ('reported', 0.027), ('querying', 0.027), ('fit', 0.027), ('aims', 0.027), ('literature', 0.027), ('first', 0.026), ('tuning', 0.025), ('modifying', 0.025), ('engine', 0.025), ('formulated', 0.024), ('majority', 0.024), ('ith', 0.024), ('exploits', 0.024), ('field', 0.024), ('defined', 0.024), ('accordingly', 0.024), ('color', 0.023), ('iterations', 0.023), ('clearly', 0.023), ('worth', 0.023), ('seven', 0.023), ('york', 0.022)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.99999976 85 nips-2004-Instance-Based Relevance Feedback for Image Retrieval

Author: Giorgio Gia\-cin\-to, Fabio Roli

2 0.21607007 134 nips-2004-Object Classification from a Single Example Utilizing Class Relevance Metrics

Author: Michael Fink

Abstract: We describe a framework for learning an object classiﬁer from a single example. This goal is achieved by emphasizing the relevant dimensions for classiﬁcation using available examples of related classes. Learning to accurately classify objects from a single training example is often unfeasible due to overﬁtting effects. However, if the instance representation provides that the distance between each two instances of the same class is smaller than the distance between any two instances from different classes, then a nearest neighbor classiﬁer could achieve perfect performance with a single training example. We therefore suggest a two stage strategy. First, learn a metric over the instances that achieves the distance criterion mentioned above, from available examples of other related classes. Then, using the single examples, deﬁne a nearest neighbor classiﬁer where distance is evaluated by the learned class relevance metric. Finding a metric that emphasizes the relevant dimensions for classiﬁcation might not be possible when restricted to linear projections. We therefore make use of a kernel based metric learning algorithm. Our setting encodes object instances as sets of locality based descriptors and adopts an appropriate image kernel for the class relevance metric learning. The proposed framework for learning from a single example is demonstrated in a synthetic setting and on a character classiﬁcation task. 1

3 0.12619428 22 nips-2004-An Investigation of Practical Approximate Nearest Neighbor Algorithms

Author: Ting Liu, Andrew W. Moore, Ke Yang, Alexander G. Gray

Abstract: This paper concerns approximate nearest neighbor searching algorithms, which have become increasingly important, especially in high dimensional perception areas such as computer vision, with dozens of publications in recent years. Much of this enthusiasm is due to a successful new approximate nearest neighbor approach called Locality Sensitive Hashing (LSH). In this paper we ask the question: can earlier spatial data structure approaches to exact nearest neighbor, such as metric trees, be altered to provide approximate answers to proximity queries and if so, how? We introduce a new kind of metric tree that allows overlap: certain datapoints may appear in both the children of a parent. We also introduce new approximate k-NN search algorithms on this structure. We show why these structures should be able to exploit the same randomprojection-based approximations that LSH enjoys, but with a simpler algorithm and perhaps with greater efﬁciency. We then provide a detailed empirical evaluation on ﬁve large, high dimensional datasets which show up to 31-fold accelerations over LSH. This result holds true throughout the spectrum of approximation levels. 1

4 0.11659783 23 nips-2004-Analysis of a greedy active learning strategy

Author: Sanjoy Dasgupta

Abstract: We abstract out the core search problem of active learning schemes, to better understand the extent to which adaptive labeling can improve sample complexity. We give various upper and lower bounds on the number of labels which need to be queried, and we prove that a popular greedy active learning rule is approximately as good as any other strategy for minimizing this number of labels. 1

5 0.10953551 89 nips-2004-Joint MRI Bias Removal Using Entropy Minimization Across Images

Author: Erik G. Learned-miller, Parvez Ahammad

Abstract: The correction of bias in magnetic resonance images is an important problem in medical image processing. Most previous approaches have used a maximum likelihood method to increase the likelihood of the pixels in a single image by adaptively estimating a correction to the unknown image bias ﬁeld. The pixel likelihoods are deﬁned either in terms of a pre-existing tissue model, or non-parametrically in terms of the image’s own pixel values. In both cases, the speciﬁc location of a pixel in the image is not used to calculate the likelihoods. We suggest a new approach in which we simultaneously eliminate the bias from a set of images of the same anatomy, but from different patients. We use the statistics from the same location across different images, rather than within an image, to eliminate bias ﬁelds from all of the images simultaneously. The method builds a “multi-resolution” non-parametric tissue model conditioned on image location while eliminating the bias ﬁelds associated with the original image set. We present experiments on both synthetic and real MR data sets, and present comparisons with other methods. 1

6 0.095733352 99 nips-2004-Learning Hyper-Features for Visual Identification

7 0.088344492 156 nips-2004-Result Analysis of the NIPS 2003 Feature Selection Challenge

8 0.081388094 155 nips-2004-Responding to Modalities with Different Latencies

9 0.081113525 125 nips-2004-Multiple Relational Embedding

10 0.07815896 40 nips-2004-Common-Frame Model for Object Recognition

11 0.077690661 192 nips-2004-The power of feature clustering: An application to object detection

12 0.071653299 121 nips-2004-Modeling Nonlinear Dependencies in Natural Images using Mixture of Laplacian Distribution

13 0.070700996 13 nips-2004-A Three Tiered Approach for Articulated Object Action Modeling and Recognition

14 0.069523126 196 nips-2004-Triangle Fixing Algorithms for the Metric Nearness Problem

15 0.064300314 111 nips-2004-Maximal Margin Labeling for Multi-Topic Text Categorization

16 0.062932834 44 nips-2004-Conditional Random Fields for Object Recognition

17 0.062720113 15 nips-2004-Active Learning for Anomaly and Rare-Category Detection

18 0.05927898 68 nips-2004-Face Detection --- Efficient and Rank Deficient

19 0.055758681 98 nips-2004-Learning Gaussian Process Kernels via Hierarchical Bayes

20 0.055341657 107 nips-2004-Making Latin Manuscripts Searchable using gHMM's

similar papers computed by lsi model

lsi for this paper:

topicId topicWeight

[(0, -0.171), (1, 0.053), (2, -0.058), (3, -0.128), (4, 0.102), (5, 0.119), (6, 0.065), (7, -0.054), (8, 0.085), (9, 0.003), (10, -0.114), (11, 0.015), (12, -0.04), (13, -0.038), (14, -0.094), (15, 0.038), (16, 0.054), (17, 0.015), (18, 0.069), (19, -0.063), (20, 0.208), (21, 0.128), (22, 0.026), (23, 0.108), (24, 0.05), (25, -0.073), (26, 0.078), (27, 0.157), (28, 0.128), (29, 0.001), (30, -0.121), (31, -0.006), (32, -0.275), (33, -0.022), (34, 0.021), (35, -0.136), (36, -0.031), (37, 0.061), (38, 0.002), (39, -0.032), (40, -0.014), (41, -0.046), (42, -0.026), (43, 0.098), (44, 0.228), (45, -0.001), (46, -0.035), (47, 0.037), (48, -0.068), (49, 0.105)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.97590989 85 nips-2004-Instance-Based Relevance Feedback for Image Retrieval

Author: Giorgio Gia\-cin\-to, Fabio Roli

2 0.67701381 22 nips-2004-An Investigation of Practical Approximate Nearest Neighbor Algorithms

Author: Ting Liu, Andrew W. Moore, Ke Yang, Alexander G. Gray

3 0.65858531 134 nips-2004-Object Classification from a Single Example Utilizing Class Relevance Metrics

Author: Michael Fink

4 0.54730761 107 nips-2004-Making Latin Manuscripts Searchable using gHMM's

Author: Jaety Edwards, Yee W. Teh, Roger Bock, Michael Maire, Grace Vesom, David A. Forsyth

Abstract: We describe a method that can make a scanned, handwritten mediaeval latin manuscript accessible to full text search. A generalized HMM is ﬁtted, using transcribed latin to obtain a transition model and one example each of 22 letters to obtain an emission model. We show results for unigram, bigram and trigram models. Our method transcribes 25 pages of a manuscript of Terence with fair accuracy (75% of letters correctly transcribed). Search results are very strong; we use examples of variant spellings to demonstrate that the search respects the ink of the document. Furthermore, our model produces fair searches on a document from which we obtained no training data. 1. Intoduction There are many large corpora of handwritten scanned documents, and their number is growing rapidly. Collections range from the complete works of Mark Twain to thousands of pages of zoological notes spanning two centuries. Large scale analyses of such corpora is currently very difﬁcult, because handwriting recognition works poorly. Recently, Rath and Manmatha have demonstrated that one can use small bodies of aligned material as supervised data to train a word spotting mechanism [7]. The result can make scanned handwritten documents searchable. Current techniques assume a closed vocabulary — one can search only for words in the training set — and search for instances of whole words. This approach is particularly unattractive for an inﬂected language, because individual words can take so many forms that one is unlikely to see all in the training set. Furthermore, one would like the method used to require very little aligned training data, so that it is possible to process documents written by different scribes with little overhead. Mediaeval Latin manuscripts are a natural ﬁrst corpus for studying this problem, because there are many scanned manuscripts and because the handwriting is relatively regular. We expect the primary user need to be search over a large body of documents — to allow comparisons between documents — rather than transcription of a particular document (which is usually relatively easy to do by hand). Desirable features for a system are: First, that it use little or no aligned training data (an ideal, which we believe may be attainable, is an unsupervised learning system). Second, that one can search the document for an arbitrary string (rather than, say, only complete words that appear in the training data). This would allow a user to determine whether a document contains curious or distinctive spellings, for example (ﬁgure 7). We show that, using a statistical model based on a generalized HMM, we can search a medieval manuscript with considerable accuracy, using only one instance each of each letter in the manuscript to train the method (22 instances in total; Latin has no j, k, w, or z). Furthermore, our method allows fairly accurate transcription of the manuscript. We train our system on 22 glyphs taken from a a 12th century latin manuscript of Terence’s Comedies (obtained from a repository of over 80 scanned medieval works maintained by Oxford University [1]). We evaluate searches using a considerable portion of this manuscript aligned by hand; we then show that fair search results are available on a different manuscript (MS. Auct. D. 2. 16, Latin Gospels with beast-headed evangelist portraits made at Landvennec, Brittany, late 9th or early 10th century, from [1]) without change of letter templates. 1.1. Previous Work Handwriting recognition is a traditional problem, too well studied to review in detail here (see [6]). Typically, online handwriting recognition (where strokes can be recorded) works better than ofﬂine handwriting recognition. Handwritten digits can now be recognized with high accuracy [2, 5]. Handwritten amounts can be read with fair accuracy, which is signiﬁcantly improved if one segments the amount into digits at the same time as one recognizes it [4, 5]. Recently several authors have proposed new techniques for search and translation in this unrestricted setting. Manmatha et al [7] introduce the technique of “word spotting,” which segments text into word images, rectiﬁes the word images, and then uses an aligned training set to learn correspondences between rectiﬁed word images and strings. The method is not suitable for a heavily inﬂected language, because words take so many forms. In an inﬂected language, the natural unit to match to is a subset of a word, rather than a whole word, implying that one should segment the text into blocks — which may be smaller than words — while recognizing. Vinciarelli et al [8] introduce a method for line by line recognition based around an HMM and quite similar to techniques used in the speech recognition community. Their method uses a window that slides along the text to obtain features; this has the difﬁculty that the same window is in some places too small (and so uninformative) and in others too big (and so spans more than one letter, and is confusing). Their method requires a substantial body of aligned training data, which makes it impractical for our applications. Close in spirit to our work is the approach to machine translation of Koehn and Knight [3]. They demonstrate that the statistics of unaligned corpora may provide as powerful constraints for training models as aligned bitexts. 2. The Model Our models for both search and transcription are based on the generalized HMM and differ only in their choice of transition model. In an HMM, each hidden node ct emits a single evidence node xt . In a generalized HMM, we allow each ct to emit a series of x’s whose length is itself a random variable. In our model, the hidden nodes correspond to letters and each xt is a single column of pixels. Allowing letters to emit sets of columns lets us accomodate letter templates of variable width. In particular, this means that we can unify segmenting ink into letters and recognizing blocks of ink; ﬁgure 3 shows an example of how useful this is. 2.1. Generating a line of text Our hidden state consists of a character label c, width w and vertical position y. The statespace of c contains the characters ‘a’-‘z’, a space ‘ ’, and a special end state Ω. Let T c be the template associated with character c, Tch , Tcw be respectively the height and width of that template, and m be the height of the image. Figure 1: Left, a full page of our manuscript, a 12’th century manuscript of Terence’s Comedies obtained from [1]. Top right, a set of lines from a page from that document and bottom right, some words in higher resolution. Note: (a) the richness of page layout; (b) the clear spacing of the lines; (c) the relatively regular handwriting. Figure 2: Left, the 22 instances, one per letter, used to train our emission model. These templates are extracted by hand from the Terence document. Right, the ﬁve image channels for a single letter. Beginning at image column 1 (and assuming a dummy space before the ﬁrst character), • • • • choose character c ∼ p(c|c−1...−n ) (an n-gram letter model) choose length w ∼ Uniform(Tcw − k, Tcw + k) (for some small k) choose vertical position y ∼ Uniform(1, m − Tch ) z,y and Tch now deﬁne a bounding box b of pixels. Let i and j be indexed from the top left of that bounding box. – draw pixel (i, j) ∼ N (Tcij , σcij ) for each pixel in b – draw all pixels above and below b from background gaussian N (µ0 , σ0 ) (See 2.2 for greater detail on pixel emission model) • move to column w + 1 and repeat until we enter the end state Ω. Inference on a gHMM is a relatively straighforward business of dynamic programming. We have used unigram, bigram and trigram models, with each transition model ﬁtted using an electronic version of Caesar’s Gallic Wars, obtained from http://www.thelatinlibrary.com. We do not believe that the choice of author should signiﬁcantly affect the ﬁtted transition model — which is at the level of characters — but have not experimented with this point. The important matter is the emission model. 2.2. The Emission Model Our emission model is as follows: Given the character c and width w, we generate a template of the required length. Each pixel in this template becomes the mean of a gaussian which generates the corresponding pixel in the image. This template has a separate mean image for each pixel channel. The channels are assumed independent given the means. We train the model by cutting out by hand a single instance of each letter from our corpus (ﬁgure 2). This forms the central portion of the template. Pixels above and below this Model Perfect transcription unigram bigram trigram matching chars 21019 14603 15572 15788 substitutions 0 5487 4597 4410 insertions 0 534 541 507 deletions 0 773 718 695 Table 1: Edit distance between our transcribed Terence and the editor’s version. Note the trigram model produces signiﬁcantly fewer letter errors than the unigram model, but that the error rate is still a substantial 25%. central box are generated from a single gaussian used to model background pixels (basically white pixels). We add a third variable yt to our hidden state indicating the vertical position of the central box. However, since we are uninterested in actually recovering this variable, during inference we sum it out of the model. The width of a character is constrained to be close to the width (tw ) of our hand cut example by setting p(w|c) = 0 for w < tw − k and w > tw + k. Here k is a small, user deﬁned integer. Within this range, p(w|c) is distributed uniformly, larger templates are created by appending pixels from the background model to the template and smaller ones by simply removing the right k-most columns of the hand cut example. For features, we generate ﬁve image representations, shown in ﬁgure 2. The ﬁrst is a grayscale version of the original color image. The second and third are generated by convolving the grayscale image with a vertical derivative of gaussian ﬁlter, separating the positive and negative components of this response, and smoothing each of these gradient images separately. The fourth and ﬁfth are generated similarly but with a horizontal derivative of gaussian ﬁlter. We have experimented with different weightings of these 5 channels. In practice we use the gray scale channel and the horizontal gradient channels. We emphasize the horizontal pieces since these seem the more discriminative. 2.3. Transcription For transcription, we model letters as coming from an n-gram language model, with no dependencies between words. Thus, the probability of a letter depends on the k letters before it, where k = n unless this would cross a word boundary in which case the history terminates at this boundary. We chose not to model word to word transition probabilities since, unlike in English, word order in Latin is highly arbitrary. This transition model is ﬁt from a corpus of ascii encoded latin. We have experimented with unigram (i.e. uniform transition probabilities), bigram and trigram letter models. We can perform transcription by ﬁtting the maximum likelihood path through any given line. Some results of this technique are shown in ﬁgure 3. 2.4. Search For search, we rank lines by the probability that they contain our search word. We set up a ﬁnite state machine like that in ﬁgure 4. In this ﬁgure, ‘bg’ represents our background model for that portion of the line not generated by our search word. We can use any of the n-gram letter models described for transcription as the transition model for ‘bg’. The probability that the line contains the search word is the probability that this FSM takes path 1. We use this FSM as the transition model for our gHMM, and output the posterior probability of the two arrows leading into the end state. 1 and 2 are user deﬁned weights, but in practice the algorithm does not appear to be particular sensitive to the choice of these parameters. The results presented here use the unigram model. Editorial translation Orator ad vos venio ornatu prologi: unigram b u rt o r a d u o s u em o o r n a t u p r o l o g r b u rt o r a d v o s v em o o r u a t u p r o l o g r fo r a t o r a d v o s v en i o o r n a t u p r o l o g i bigram trigram Figure 3: We transcribe the text by ﬁnding the maximum likelihood path through the gHMM. The top line shows the standard version of the line (obtained by consensus among editors who have consulted various manuscripts; we obtained this information in electronic form from http://www.thelatinlibrary.com). Below, we show the line as segmented and transcribed by unigram, bigram and trigram models; the unigram and bigram models transcribe one word as “vemo”, but the stronger trigram model forces the two letters to be segmented and correctly transcribes the word as “venio”, illustrating the considerable beneﬁt to be obtained by segmenting only at recognition time. 1 − ε1 Path 1 1 − ε2 a b bg ε1 Ω bg Path 2 ε2 Figure 4: The ﬁnite state machine to search for the word ‘ab.’ ‘bg’ is a place holder for the larger ﬁnite state machine deﬁned by our language model’s transition matrix. 3. Results Figure 1 shows a page from our collection. This is a scanned 12th century manuscript of Terence’s Comedies, obtained from the collection at [1]. In preprocessing, we extract individual lines of text by rotating the image to various degrees and projecting the sum of the pixel values onto the y-axis. We choose the orientation whose projection vector has the lowest entropy, and then segment lines by cutting at minima of this projection. Transcription is not our primary task, but methods that produce good transcriptions are going to support good searches. The gHMM can produce a surprisingly good transcription, given how little training data is used to train the emission model. We aligned an editors version of Terence with 25 pages from the manuscript by hand, and computed the edit distance between the transcribed text and the aligned text; as table 1 indicates, approximately 75% of letters are read correctly. Search results are strong. We show results for two documents. The ﬁrst set of results refers to the edition of Terence’s Comedies, from which we took the 22 letter instances. In particular, for any given search term, our process ranks the complete set of lines. We used a hand alignment of the manuscript to determine which lines contained each term; ﬁgure 5 shows an overview of searches performed using every word that appears in the 50 100 150 200 250 300 350 400 450 500 550 Figure 5: Our search ranks 587 manuscript lines, with higher ranking lines more likely to contain the relevant term. This ﬁgure shows complete search results for each term that appears more than three times in the 587 lines. Each row represents the ranked search results for a term, and a black mark appears if the search term is actually in the line; a successful search will therefore appear as a row which is wholly dark to the left, and then wholly light. All 587 lines are represented. More common terms are represented by lower rows. More detailed results appear in ﬁgure 5 and ﬁgure 6; this summary ﬁgure suggests almost all searches are highly successful. document more than three times, in particular, showing which of the ranked set of lines actually contained the search term. For almost every search, the term appears mainly in the lines with higher rank. Figure 6 contains more detailed information for a smaller set of words. We do not score the position of a word in a line (for practical reasons). Figure 7 demonstrates (a) that our search respects the ink of the document and (b) that for the Terence document, word positions are accurately estimated. The spelling of mediaeval documents is typically cleaned up by editors; in our manuscript, the scribe reliably spells “michi” for the standard “mihi”. A search on “michi” produces many instances; a search on “mihi” produces none, because the ink doesn’t have any. Notice this phenomenon also in the bottom right line of ﬁgure 7, the scribe writes “habet, ut consumat nunc cum nichil obsint doli” and the editor gives “habet, ut consumat nunc quom nil obsint doli.” Figure 8 shows that searches on short strings produce many words containing that string as one would wish. 4. Discussion We have shown that it is possible to make at least some handwritten mediaeval manuscripts accessible to full text search, without requiring an aligned text or much supervisory data. Our documents have very regular letters, and letter frequencies — which can be obtained from transcribed Latin — appear to provide so powerful a cue that relatively little detailed information about letter shapes is required. Linking letter segmentation and recognition has thoroughly beneﬁcial effects. This suggests that the pool of manuscripts that can be made accessible in this way is large. In particular, we have used our method, trained on 22 instances of letters from one document, to search another document. Figure 9 shows the results from two searches of our second document (MS. Auct. D. 2. 16, Latin Gospels with beast-headed evangelist portraits made at Landvennec, Brittany, late 9th or early 10th century, from [1]). No information from this document was used in training at all; but letter 1tu arbitror pater etiam nisi factum primum siet vero illi inter hic michi ibi qui tu ibi michi 0.9 0.8 0.7 qui hic 0.6 inter 0.5 illi 0.4 siet 0.3 vero 0.2 nisi 0.1 50 100 150 200 250 300 350 400 450 500 550 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Figure 6: On the left, search results for selected words (indicated on the leftmost column). Each row represents the ranked search results for a term, and a black mark appears if the search term is actually in the line; a successful search will therefore appear as a row which is wholly dark to the left, and then wholly light. Note only the top 300 results are represented, and that lines containing the search term are almost always at or close to the top of the search results (black marks to the left). On the right, we plot precision against recall for a set of different words by taking the top 10, 20, ... lines returned from the search, and checking them against the aligned manuscript. Note that, once all cases have been found, if the size of the pool is increased the precision will fall with 100% recall; many words work well, with most of the ﬁrst 20 or so lines returned containing the search term. shapes are sufﬁciently well shared that the search is still useful. All this suggests that one might be able to use EM to link three processes: one that clusters to determine letter shapes; one that segments letters; and one that imposes a language model. Such a system might be able to make handwritten Latin searchable with no training data. References [1] Early Manuscripts at Oxford University. Bodleian library ms. auct. f. 2.13. http://image.ox.ac.uk/. [2] Serge Belongie, Jitendra Malik, and Jan Puzicha. Shape matching and object recognition using shape contexts. IEEE T. Pattern Analysis and Machine Intelligence, 24(4):509–522, 2002. [3] Philipp Koehn and Kevin Knight. Estimating word translation probabilities from unrelated monolingual corpora. In Proc. of the 17th National Conf. on AI, pages 711–715. AAAI Press / The MIT Press, 2000. [4] Y. LeCun, L. Bottou, and Y. Bengio. Reading checks with graph transformer networks. In International Conference on Acoustics, Speech, and Signal Processing, volume 1, pages 151–154, Munich, 1997. IEEE. [5] Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324, 1998. [6] R. Plamondon and S.N. Srihari. Online and off-line handwriting recognition: a comprehensive survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(1):63–84, 2000. [7] T. M. Rath and R. Manmatha. Word image matching using dynamic time warping. In Proc. of the Conf. on Computer Vision and Pattern Recognition (CVPR), volume 2, pages 521–527, 2003. [8] Alessandro Vinciarelli, Samy Bengio, and Horst Bunke. Ofﬂine recognition of unconstrained handwritten texts using hmms and statistical language models. IEEE Trans. Pattern Anal. Mach. Intell., 26(6):709–720, 2004. michi: Spe incerta certum mihi laborem sustuli, mihi: Faciuntne intellegendo ut nil intellegant? michi: Nonnumquam conlacrumabat. placuit tum id mihi. mihi: Placuit: despondi. hic nuptiis dictust dies. michi: Sto exspectans siquid mi imperent. venit una,

5 0.46850002 196 nips-2004-Triangle Fixing Algorithms for the Metric Nearness Problem

Author: Suvrit Sra, Joel Tropp, Inderjit S. Dhillon

Abstract: Various problems in machine learning, databases, and statistics involve pairwise distances among a set of objects. It is often desirable for these distances to satisfy the properties of a metric, especially the triangle inequality. Applications where metric data is useful include clustering, classiﬁcation, metric-based indexing, and approximation algorithms for various graph problems. This paper presents the Metric Nearness Problem: Given a dissimilarity matrix, ﬁnd the “nearest” matrix of distances that satisfy the triangle inequalities. For p nearness measures, this paper develops efﬁcient triangle ﬁxing algorithms that compute globally optimal solutions by exploiting the inherent structure of the problem. Empirically, the algorithms have time and storage costs that are linear in the number of triangle constraints. The methods can also be easily parallelized for additional speed. 1

6 0.44578788 23 nips-2004-Analysis of a greedy active learning strategy

7 0.44250193 89 nips-2004-Joint MRI Bias Removal Using Entropy Minimization Across Images

8 0.43503019 192 nips-2004-The power of feature clustering: An application to object detection

9 0.3862555 15 nips-2004-Active Learning for Anomaly and Rare-Category Detection

10 0.36298734 199 nips-2004-Using Machine Learning to Break Visual Human Interaction Proofs (HIPs)

11 0.35823339 155 nips-2004-Responding to Modalities with Different Latencies

12 0.35475692 99 nips-2004-Learning Hyper-Features for Visual Identification

13 0.35334361 40 nips-2004-Common-Frame Model for Object Recognition

14 0.33950582 14 nips-2004-A Topographic Support Vector Machine: Classification Using Local Label Configurations

15 0.33870587 25 nips-2004-Assignment of Multiplicative Mixtures in Natural Images

16 0.3328487 191 nips-2004-The Variational Ising Classifier (VIC) Algorithm for Coherently Contaminated Data

17 0.32996365 127 nips-2004-Neighbourhood Components Analysis

18 0.32837906 156 nips-2004-Result Analysis of the NIPS 2003 Feature Selection Challenge

19 0.31370711 68 nips-2004-Face Detection --- Efficient and Rank Deficient

20 0.30699593 109 nips-2004-Mass Meta-analysis in Talairach Space

similar papers computed by lda model

lda for this paper:

topicId topicWeight

[(13, 0.644), (15, 0.085), (26, 0.03), (31, 0.011), (33, 0.092), (35, 0.016), (50, 0.015)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.97006184 85 nips-2004-Instance-Based Relevance Feedback for Image Retrieval

Author: Giorgio Gia\-cin\-to, Fabio Roli

2 0.96079296 27 nips-2004-Bayesian Regularization and Nonnegative Deconvolution for Time Delay Estimation

Author: Yuanqing Lin, Daniel D. Lee

Abstract: Bayesian Regularization and Nonnegative Deconvolution (BRAND) is proposed for estimating time delays of acoustic signals in reverberant environments. Sparsity of the nonnegative ﬁlter coefﬁcients is enforced using an L1 -norm regularization. A probabilistic generative model is used to simultaneously estimate the regularization parameters and ﬁlter coefﬁcients from the signal data. Iterative update rules are derived under a Bayesian framework using the Expectation-Maximization procedure. The resulting time delay estimation algorithm is demonstrated on noisy acoustic data.

3 0.95083755 176 nips-2004-Sub-Microwatt Analog VLSI Support Vector Machine for Pattern Classification and Sequence Estimation

Author: Shantanu Chakrabartty, Gert Cauwenberghs

Abstract: An analog system-on-chip for kernel-based pattern classiﬁcation and sequence estimation is presented. State transition probabilities conditioned on input data are generated by an integrated support vector machine. Dot product based kernels and support vector coefﬁcients are implemented in analog programmable ﬂoating gate translinear circuits, and probabilities are propagated and normalized using sub-threshold current-mode circuits. A 14-input, 24-state, and 720-support vector forward decoding kernel machine is integrated on a 3mm×3mm chip in 0.5µm CMOS technology. Experiments with the processor trained for speaker veriﬁcation and phoneme sequence estimation demonstrate real-time recognition accuracy at par with ﬂoating-point software, at sub-microwatt power. 1

4 0.9198097 138 nips-2004-Online Bounds for Bayesian Algorithms

Author: Sham M. Kakade, Andrew Y. Ng

Abstract: We present a competitive analysis of Bayesian learning algorithms in the online learning setting and show that many simple Bayesian algorithms (such as Gaussian linear regression and Bayesian logistic regression) perform favorably when compared, in retrospect, to the single best model in the model class. The analysis does not assume that the Bayesian algorithms’ modeling assumptions are “correct,” and our bounds hold even if the data is adversarially chosen. For Gaussian linear regression (using logloss), our error bounds are comparable to the best bounds in the online learning literature, and we also provide a lower bound showing that Gaussian linear regression is optimal in a certain worst case sense. We also give bounds for some widely used maximum a posteriori (MAP) estimation algorithms, including regularized logistic regression. 1

5 0.91591185 114 nips-2004-Maximum Likelihood Estimation of Intrinsic Dimension

Author: Elizaveta Levina, Peter J. Bickel

Abstract: We propose a new method for estimating intrinsic dimension of a dataset derived by applying the principle of maximum likelihood to the distances between close neighbors. We derive the estimator by a Poisson process approximation, assess its bias and variance theoretically and by simulations, and apply it to a number of simulated and real datasets. We also show it has the best overall performance compared with two other intrinsic dimension estimators. 1

6 0.8666243 36 nips-2004-Class-size Independent Generalization Analsysis of Some Discriminative Multi-Category Classification

7 0.81611347 5 nips-2004-A Harmonic Excitation State-Space Approach to Blind Separation of Speech

8 0.68400967 181 nips-2004-Synergies between Intrinsic and Synaptic Plasticity in Individual Model Neurons

9 0.62998754 135 nips-2004-On-Chip Compensation of Device-Mismatch Effects in Analog VLSI Neural Networks

10 0.62164831 28 nips-2004-Bayesian inference in spiking neurons

11 0.61165482 58 nips-2004-Edge of Chaos Computation in Mixed-Mode VLSI - A Hard Liquid

12 0.60220408 22 nips-2004-An Investigation of Practical Approximate Nearest Neighbor Algorithms

13 0.59922153 131 nips-2004-Non-Local Manifold Tangent Learning

14 0.59659159 163 nips-2004-Semi-parametric Exponential Family PCA

15 0.59411991 102 nips-2004-Learning first-order Markov models for control

16 0.59397507 116 nips-2004-Message Errors in Belief Propagation

17 0.59310019 75 nips-2004-Heuristics for Ordering Cue Search in Decision Making

18 0.58879745 204 nips-2004-Variational Minimax Estimation of Discrete Distributions under KL Loss

19 0.58588469 119 nips-2004-Mistake Bounds for Maximum Entropy Discrimination

20 0.58345723 152 nips-2004-Real-Time Pitch Determination of One or More Voices by Nonnegative Matrix Factorization