cvpr cvpr2013 cvpr2013-69 cvpr2013-69-reference knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Tomasz Trzcinski, Mario Christoudias, Pascal Fua, Vincent Lepetit
Abstract: Binary keypoint descriptors provide an efficient alternative to their floating-point competitors as they enable faster processing while requiring less memory. In this paper, we propose a novel framework to learn an extremely compact binary descriptor we call BinBoost that is very robust to illumination and viewpoint changes. Each bit of our descriptor is computed with a boosted binary hash function, and we show how to efficiently optimize the different hash functions so that they complement each other, which is key to compactness and robustness. The hash functions rely on weak learners that are applied directly to the imagepatches, whichfrees usfrom any intermediate representation and lets us automatically learn the image gradient pooling configuration of the final descriptor. Our resulting descriptor significantly outperforms the state-of-the-art binary descriptors and performs similarly to the best floating-point descriptors at a fraction of the matching time and memory footprint.
[1] K. Ali, F. Fleuret, D. Hasler, and P. Fua. A Real-Time Deformable Detector. PAMI, 34(2):225–239, 2012. 4
[2] H. Bay, T. Tuytelaars, and L. Van Gool. SURF: Speeded Up Robust Features. In ECCV’06. 1, 5 222888778088 Train: Yosemite, Test: Notre Dame False Positive Rate Figure 8. Our BinBoost descriptor’s performance compared with different binarization methods applied on L-BGM. Binarizing the discriminative projections found with L-BGM either by simple thresholding or with Iterative Quantization (ITQ) results in large binarization errors significantly reducing its accuracy. On the other hand, the sequential projection learning of S3PLH requires a fairly large number of bits to recover L-BGM’s original performance. In contrast, by jointly optimizing over the feature weighting and pooling strategy of each bit, our BinBoost approach results in a highly compact and accurate binary descriptor whose performance is similar with L-BGM but at a fraction of the storage cost.
[3] M. Brown, G. Hua, and S. Winder. Discriminative Learning of Local Image Descriptors. PAMI, 2011. 1, 2, 4, 5, 6, 7
[4] M. Calonder, V. Lepetit, M. Ozuysal, T. Trzcinski, C. Strecha, and P. Fua. BRIEF: Computing a Local Binary Descriptor Very Fast. PAMI, 34(7):1281–1298, 2012. 1, 2, 6
[5] V. Chandrasekhar, G. Takacs, D. Chen, S. Tsai, R. Grzeszczuk, and B. Girod. CHoG: Compressed Histogram of Gradients a Low Bit-Rate Feature Descriptor. In CVPR’09. 2
[6] O. Chapelle, P. Shivaswamy, S. Vadrevu, K. Weinberger, Y. Zhang, and B. Tseng. Boosted Multi-Task Learning. Machine Learning, 2010. 3
[7] P. Doll a´r, Z. Tu, P. Perona, and S. Belongie. Integral Channel Features. In BMVC’09. 3
[8] Y. Freund and R. Schapire. A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. In European Conference on Computational Learning Theory, 1995. 3
[9] Y. Gong, S. Lazebnik, A. Gordo, and F. Perronnin. Iterative Quantization: A Procrustean Approach to Learning Binary Codes for Large-Scale Image Retrieval. PAMI, 2012. 1, 2, 6, 7
[10] K. Grauman and T. Darrell. The Pyramid Match Kernel: Discriminative Classification with Sets of Image Features. In ICCV’05. 2
[11] T. Hastie, R. Tibshirani, and J. Friedman. The Elements of
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25] Statistical Learning. 2001 . 4 P. Jain, B. Kulis, J. Davis, and I. Dhillon. Metric and Kernel Learning Using a Linear Transformation. JMLR, 2012. 2 H. J ´egou, M. Douze, C. Schmid, and P. P ´erez. Aggregating Local Descriptors into a Compact Image Representation. In CVPR’10. 2 B. Kulis and T. Darrell. Learning to Hash with Binary Reconstructive Embeddings. In NIPS’09. 2 S. Leutenegger, M. Chli, and R. Siegwart. BRISK: Binary Robust Invariant Scalable Keypoints. In ICCV’11. 1, 2, 6 W. Liu, J. Wang, R. Ji, Y.-G. Jiang, and S.-F. Chang. Supervised Hashing with Kernels. In CVPR ’12. 2, 3, 4 D. Lowe. Distinctive Image Features from Scale-Invariant Keypoints. IJCV, 20(2):91–1 10, 2004. 1, 5 M. Malekesmaeili, R. Ward, and M. Fatourechi. A Fast Approximate Nearest Neighbor Search Algorithm in the Hamming Space. PAMI, 2012. 1 M. Norouzi and D. Fleet. Minimal Loss Hashing for Compact Binary Codes. In ICML’11. 2 M. Norouzi, A. Punjani, and D. Fleet. Fast Search in Hamming Space with Multi-Index Hashing. In CVPR ’12. 1 S. Rosset, J. Zhu, and T. Hastie. Boosting as a Regularized Path to a Maximum Margin Classifier. JMLR, 2004. 3 E. Rublee, V. Rabaud, K. Konolidge, and G. Bradski. ORB: An Efficient Alternative to SIFT or SURF. In ICCV’11. 1, 2 R. Salakhutdinov and G. Hinton. Semantic Hashing. International Journal of Approximate Reasoning, 2009. 2, 3 R. E. Schapire and Y. Singer. Improved Boosting Algorithms Using Confidence Rated Predictions. Machine Learning, 1999. 4 G. Shakhnarovich. Learning Task-Specific Similarity. PhD thesis, 2006. 3, 4, 6
[26] K. Simonyan, A. Vedaldi, and A. Zisserman. Descriptor Learning Using Convex Optimisation. In ECCV’12. 2, 6, 7
[27] C. Strecha, A. Bronstein, M. Bronstein, and P. Fua. LDAHash: Improved Matching with Smaller Descriptors. PAMI, 34(1), 2012. 1, 2, 5, 7
[28] A. Torralba, R. Fergus, and Y. Weiss. Small Codes and Large Databases for Recognition. In CVPR ’08. 3
[29] T. Trzcinski, M. Christoudias, V. Lepetit, and P. Fua. Learning Image Descriptors with the Boosting-Trick. In NIPS’12. 2, 3, 4, 6, 7
[30] T. Trzcinski and V. Lepetit. Efficient Discriminative Projections for Compact Binary Descriptors. In ECCV’12. 1, 2 [3 1] A. Vedaldi. http : / /www .vlfeat . org/ ˜vedaldi / code / s i ftpp .html . 6, 7
[32] P. Viola and M. Jones. Rapid Object Detection Using a Boosted Cascade of Simple Features. In CVPR ’01. 3
[33] G. Wang and Q. Wu. Quasi-Perspective Projection Model: Theory and Application to Structure and Motion Factorization from Uncalibrated Image Sequences. IJCV, 87(3):213– 234, 2010.
[34] J. Wang, S. Kumar, and S.-F. Chang. Sequential Projection Learning for Hashing with Compact Codes. In ICML’10. 2, 3, 4, 7
[35] Y. Weiss, R. Fergus, and A. Torralba. Multidimensional Spectral Hashing. In ECCV’12. 2
[36] Y. Weiss, A. Torralba, and R. Fergus. Spectral Hashing. NIPS, 21: 1753–1760, 2009. 2
[37] C. Zitnick. Binary Coherent Edge Descriptors. In ECCV’10. 2 222888778199