Progressive Bi-C3D Pose Grammar for Human Pose Estimation
Authors: Lu Zhou, Yingying Chen, Jinqiao Wang, Hanqing Lu13033-13040
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Without bells and whistles, our method achieves competitive performance on both MPII and LSP benchmarks compared with previous methods, which confirms the feasibility and effectiveness of C3D in information interactions. |
| Researcher Affiliation | Academia | 1National Lab of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing, China 2University of Chinese Academy of Sciences, Beijing, China |
| Pseudocode | No | The paper describes mathematical formulations for Conv LSTM in Section 3.5, but it does not present structured pseudocode or an algorithm block. |
| Open Source Code | No | The paper does not contain any explicit statement or link indicating that the source code for the described methodology is publicly available. |
| Open Datasets | Yes | The experiments are carried on two widely applied benchmarks MPII (Andriluka et al. 2014) and LSP (Johnson and Everingham 2010). |
| Dataset Splits | Yes | We conduct our experiments on the 25925 training images and 2958 valid images. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU models, CPU types) used for running the experiments, only stating 'We conduct all our experiments on the platform of Pytorch'. |
| Software Dependencies | No | The paper mentions using 'Pytorch' but does not provide specific version numbers for any software dependencies. |
| Experiment Setup | Yes | Learning rate is set as 5e-4 at the beginning and dropped by 10 at 150 epoch and 170 epoch respectively. We utilize the RMSprop (Tieleman and Hinton 2012) algorithm to update the parameters of the model. We crop the image to the size of 256 256 and person expected to be estimated is located at the center of the cropped patch with roughly the same scale. We rotate the cropped patch by 30 and scale the image by a random number. Random color jittering, shearing and flipping are involved as well. Six-scale (0.8,0.9,1.0,1.1,1.2,1.3) image pyramids combined with flipping are adopted during testing. The grammar module is appended at the end of the eight-stack hourglass model. |