Discriminatively learning for representing local image features with quadruplet model

Abstract

Traditional hand-crafted features for representing local image patches are evolving into current data-driven and learning-based image feature, but learning a robust and discriminative descriptor which is capable of controlling various patch-level computer vision tasks is still an open problem. In this work, we propose a novel deep convolutional neural network (CNN) to learn local feature descriptors. We utilize the quadruplets with positive and negative training samples, together with a constraint to restrict the intra-class variance, to learn good discriminative CNN representations. Compared with previous works, our model reduces the overlap in feature space between corresponding and non-corresponding patch pairs, and mitigates margin varying problem caused by commonly used triplet loss. We demonstrate that our method achieves better embedding result than some latest works, like PN-Net and TN-TG, on benchmark dataset.

References

[1] MoltonN., DavisonA. J., ReidI.. Locally Planar Patch Features for Real-Time Structure from Motion, 2004, 1

[2] SeitzS. M., CurlessB., DiebelJ., ScharsteinD., SzeliskiR.. A Comparison and Evaluation of Multi-View Stereo Reconstruction Algorithms, 2006, 519

[3] SzeliskiR.. Foundations and Trends in Computer Graphics and Vision, 2006, 2: 1

[4] LoweD. G.. International Journal of Computer Vision, 2004, 60: 91

[5] BayH., EssA., TuytelaarsT., Van GoolL.. Computer Vision and Image Understanding, 2008, 110: 346

[6] Simo-SerraE., TrullsE., FerrazL., KokkinosI., FnaP., Moreno-NoguerF.. Discriminative Learning of Deep Convolutional Feature Point Descriptors, 2015, 118

[7] ZagoruykoS., KomodakisN.. Learning to Compare Image Patches via Convolutional Neural Networks, 2015, 4353

[8] BalntasV., JohnsE., TangL., MikolajczykK.. PN-Net: Conjoined Triple Deep Network for Learning Local Image Descriptors, 2016,

[9] KumarB. G. V., CarneiroG., ReidI.. Learning Local Image Descriptors with Deep Siamese and Triplet Convolutional Networks by minimising global loss Functions, 2016, 5385

[10] JiaY., ShelhamerE., DonahueJ., KarayevS., LongJ., GirshickR., GuadarramaS., DarrellT.. Caffe: Convolutional Architecture for Fast Feature Embedding, 2014,

[11] BrownM., HuaG., WinderS.. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33: 43

[12] SimonyanK., VedaldiA., ZissermanA.. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36: 1573

This work has been supported by the Natural Science Foundation of Zhejiang Province (No.Y16F020023). This paper was presented in part at the CCF Chinese Conference on Computer Vision, Tianjin, 2017. This paper was recommended by the program committee.