Local Similarity-Aware Deep Feature Embedding
Authors: Chen Huang, Chen Change Loy, Xiaoou Tang
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our local similarity-aware feature embedding not only demonstrates faster convergence and boosted performance on two complex image retrieval datasets, its large margin nature also leads to superior generalization results under the large and open set scenarios of transfer learning and zero-shot learning on Image Net 2010 and Image Net-10K datasets. |
| Researcher Affiliation | Academia | Chen Huang Chen Change Loy Xiaoou Tang Department of Information Engineering, The Chinese University of Hong Kong {chuang,ccloy,xtang}@ie.cuhk.edu.hk |
| Pseudocode | No | The paper includes diagrams of network architecture (Figure 2) but does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide an explicit statement or link to open-source code for the described methodology. |
| Open Datasets | Yes | Specifically, we use the CUB-200-2011 [32] dataset with 200 bird classes and 11,788 images. We employ the first 100 classes (5,864 images) for training, and the remaining 100 classes (5,924 images) for testing. Another used dataset is CARS196 [15] with 196 car classes and 16,185 images. The first 98 classes (8,054 images) are used for training, and the other 98 classes are retained for testing (8,131 images). |
| Dataset Splits | No | The paper specifies training and testing splits for the datasets but does not explicitly mention a separate validation split with specific sizes or percentages. |
| Hardware Specification | Yes | Figure 4-(top) shows that the proposed PDDM leads to 2 faster convergence in 200 epochs (28 hours on a Titan X GPU) and lower converged loss than the regular Euclidean metric |
| Software Dependencies | No | The paper mentions using "Goog Le Net [31]" and "Caffe Net [16]" but does not provide specific version numbers for these or other software dependencies. |
| Experiment Setup | Yes | For all experiments, we choose by grid search the mini-batch size m = 64, initial learning rate 1 10 4, momentum 0.9, margin parameters α = 0.5, β = 1 in Eqs. (3, 4), and regularization parameters λ = 0.5, γ = 5 10 4 (λ balances the metric loss Em against the embedding loss Ee). The entire network is trained for a maximum of 400 epochs until convergence. |