Discriminative Semi-Coupled Projective Dictionary Learning for Low-Resolution Person Re-Identification
Authors: Kai Li, Zhengming Ding, Sheng Li, Yun Fu
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on three public datasets show the superiority of the proposed method to the state-of-the-art ones. |
| Researcher Affiliation | Collaboration | Department of Electrical & Computer Engineering, Northeastern University, Boston, USA Adobe Research, USA College of Computer & Information Science, Northeastern University, Boston, USA |
| Pseudocode | Yes | Algorithm 1: DSPDL for person re-ID. |
| Open Source Code | No | The paper does not provide an explicit statement or link for open-source code for the described methodology. |
| Open Datasets | Yes | We employed three datasets for performance evaluation: the VIPe R dataset (Gray and Tao 2008), the CUHK01 dataset (Zhao, Ouyang, and Wang 2014), and the QMULi LIDS dataset (Zheng, Gong, and Xiang 2009). |
| Dataset Splits | No | The paper mentions splitting data into training and testing parts but does not specify a separate validation split: 'all the pedestrian pairs were randomly divided into two equal parts, with one part serving for training and the other for testing.' |
| Hardware Specification | No | The paper does not specify the hardware (e.g., GPU models, CPU types) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific version numbers for software dependencies or libraries used in the implementation. |
| Experiment Setup | Yes | There are two major parameters in our algorithm: λ1 and λ2. λ2 controls the scale of a variable; thus we set it empirically as a small value λ2 = 0.01. ... DSPDL reaches the best performances when λ1 = 0.1. Thus, λ1 was set as 0.1 in our method as default. ... we down-sampled all images from one view with the rate 1/8, and kept images from the other view unchanged to simulate the great resolution differences. |