Similarity-Preserving Binary Signature for Linear Subspaces
Authors: Jianqiu Ji, Jianmin Li, Shuicheng Yan, Qi Tian, Bo Zhang
AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on face recognition demonstrate the effectiveness of the binary signature in terms of recognition accuracy, speed and storage requirement. The results show that, compared with the exact method, the approximation with the binary signatures achieves an order of magnitude speedup, while requiring significantly smaller amount of storage space, yet it still accurately preserves the similarity, and achieves high recognition accuracy comparable to the exact method in face recognition. |
| Researcher Affiliation | Academia | State Key Laboratory of Intelligent Technology and Systems, Tsinghua National Laboratory for Information Science and Technology (TNList), Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China jijq10@mails.tsinghua.edu.cn, {lijianmin, dcszb}@mail.tsinghua.edu.cn Department of Electrical and Computer Engineering, National University of Singapore, Singapore, 117576 eleyans@nus.edu.sg Department of Computer Science, University of Texas at San Antonio, qi.tian@utsa.edu |
| Pseudocode | No | The paper does not contain any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not mention providing open-source code for the described methodology. |
| Open Datasets | Yes | Datasets: we use the union of two face datasets, the Extended Yale Face Database B (Georghiades, Belhumeur, and Kriegman 2001)(Lee, Ho, and Kriegman 2005) and the PIE database (Sim, Baker, and Bsat 2002). Note that both of these datasets are the processed versions1 (He et al. 2005)(Cai et al. 2006)(Cai et al. 2007)(Cai, He, and Han 2007). 1http://www.cad.zju.edu.cn/home/dengcai/Data/Face Data.html |
| Dataset Splits | No | The two datasets have been split into training sets and testing sets respectively. The training set of the Extended Yale Face Database B contains around 50 images per individual, and the rest (around 14) images are in the testing set. The training set of the PIE database has about 20 images per individual, and the rest (around 14) images are in the testing set. The paper specifies training and testing splits, but no explicit validation split. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory specifications) used for running the experiments. |
| Software Dependencies | No | The paper does not list specific software dependencies with version numbers. |
| Experiment Setup | Yes | Experiment Setup: each face image is vectorized to a 1024 x 1 vector. Several vectors of the same individual constitute a subspace. For each of these subspaces, we fit a subspace of dimension 9 by taking the first 9 principal components, and use this subspace to represent each individual in the training set. The same goes for the testing set, except that we test query subspaces with dimensions dq = 4, 9, 13 respectively. Thus there are 106 subspaces in the training set and testing set respectively. ... We test different lengths K of the binary signatures ranging from 500 to 3000. |