Relationship-Aware Hard Negative Generation in Deep Metric Learning

Data relationships and the impact of synthetic loss have not been concerned by previous sample generation methods, which lead to bias in model training. To address above problem, in this paper, we propose a relationship-aware hard negative generation (RHNG) method. First, we build a global minimum spanning tree for all categories to measure the data distribution, which is used to constrain hard sample generation. Second, we construct a dynamic weight parameter which reflects the convergence of the model to guide the synthetic loss to train the model. Experimental results show that the proposed method outperforms the state-of-the-art methods in terms of retrieval and clustering tasks.

[1]  Jian Wang,et al.  Deep Metric Learning with Angular Loss , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[2]  Silvio Savarese,et al.  Deep Metric Learning via Lifted Structured Feature Embedding , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[3]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[4]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Wei-Shi Zheng,et al.  Cross-View Asymmetric Metric Learning for Unsupervised Person Re-Identification , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[6]  Kihyuk Sohn,et al.  Improved Deep Metric Learning with Multi-class N-pair Loss Objective , 2016, NIPS.

[7]  Xudong Lin,et al.  Deep Adversarial Metric Learning , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[8]  James Philbin,et al.  FaceNet: A unified embedding for face recognition and clustering , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  Wenge Rong,et al.  Connecting Factorization and Distance Metric Learning for Social Recommendations , 2017, KSEM.

[10]  Vincent Lepetit,et al.  3D Pose Estimation and 3D Model Retrieval for Objects in the Wild , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[11]  Kilian Q. Weinberger,et al.  Distance Metric Learning for Large Margin Nearest Neighbor Classification , 2005, NIPS.

[12]  R. Prim.  Shortest connection networks and some generalizations , 1957 .

[13]  Rui Yu,et al.  Hard-Aware Point-to-Set Deep Metric for Person Re-identification , 2018, ECCV.

[14]  J. Stuart Hunter,et al.  The exponentially weighted moving average , 1986 .

[15]  Pietro Perona,et al.  The Caltech-UCSD Birds-200-2011 Dataset , 2011 .

[16]  Jonathan Krause,et al.  3D Object Representations for Fine-Grained Categorization , 2013, 2013 IEEE International Conference on Computer Vision Workshops.

[17]  Jie Li,et al.  Unsupervised Semantic-Preserving Adversarial Hashing for Image Search , 2019, IEEE Transactions on Image Processing.

[18]  Chen Huang,et al.  Local Similarity-Aware Deep Feature Embedding , 2016, NIPS.

[19]  Bo Yang,et al.  Trust-Distrust Aware Recommendation by Integrating Metric Learning with Matrix Factorization , 2018, KSEM.

[20]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[21]  Hongtao Lu,et al.  An Adversarial Approach to Hard Triplet Generation , 2018, ECCV.

[22]  Jiwen Lu,et al.  Hardness-Aware Deep Metric Learning , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[23]  Cheng Deng,et al.  Deep Asymmetric Metric Learning via Rich Relationship Mining , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[24]  Yann LeCun,et al.  Dimensionality Reduction by Learning an Invariant Mapping , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[25]  Huchuan Lu,et al.  Stepwise Metric Promotion for Unsupervised Video Person Re-identification , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[26]  Weilin Huang,et al.  Deep Metric Learning with Hierarchical Triplet Loss , 2018, ECCV.