Sequential End-to-end Network for Efficient Person Search
Authors: Zhengjia Li, Duoqian Miao2011-2019
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on two widely used person search benchmarks, CUHK-SYSU and PRW, have shown that our method achieves state-of-the-art results. |
| Researcher Affiliation | Academia | Zhengjia Li1,2, Duoqian Miao1,2* 1Department of Computer Science and Technology, Tongji University, Shanghai 201804, China 2Key Laboratory of Embedded System and Service Computing, Ministry of Education, Shanghai 201804, China |
| Pseudocode | Yes | Algorithm 1: CBGM |
| Open Source Code | No | The paper does not provide an explicit statement about releasing the source code or a direct link to a code repository. |
| Open Datasets | Yes | CUHK-SYSU (Xiao et al. 2017) is a large scale person search dataset... PRW is another widely used dataset (Zheng et al. 2017) |
| Dataset Splits | No | The paper specifies training and test sets but does not explicitly mention or detail a separate validation set split (e.g., by size or method of creation). |
| Hardware Specification | Yes | We implement our model with Py Torch (Paszke et al. 2017) and run all experiments on one NVIDIA Tesla V100 GPU. |
| Software Dependencies | No | The paper mentions implementing the model with 'Py Torch (Paszke et al. 2017)', but it does not specify a version number for PyTorch or any other software dependencies. |
| Experiment Setup | Yes | During training, batch size is 5 and each image is resized to 900 × 1500 pixels. Our model is optimized by Stochastic Gradient Descent (SGD) for 20 epochs (18 epochs for PRW) with initial learning rate of 0.003 which is warmed up during the first epoch and decreased by 10 at the 16-th epoch. The momentum and weight decay of SGD are set to 0.9 and 5 × 10−4 individually. For CUHK-SYSU/PRW, the circular queue size of OIM is set to 5000/500. At test time, NMS with 0.4/0.5 threshold is used to remove redundant boxes detected by the first/second head. |