Pose-Assisted Multi-Camera Collaboration for Active Object Tracking

Authors: Jing Li, Jing Xu, Fangwei Zhong, Xiangyu Kong, Yu Qiao, Yizhou Wang759-766

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experimental results demonstrate that our system outperforms all the baselines and is capable of generalizing to unseen environments.
Researcher Affiliation Collaboration 1Center for Data Science, Peking University 2Computer Science Dept., Sch l of EECS, Peking University 3Advanced Innovation Center for Future Visual Entertainment(AICFVE), Beijing Film Academy 4Key Lab. of System Control and Information Processing (Mo E), Shanghai; Automation Dept., Shanghai Jiao Tong University 5Center on Frontiers of Computing Studies, Peking University 6Deepwise AI Lab ... This work was supported by MOST-2018AAA0102004, NSFC-61625201, NSFC61527804, Qualcomm University Research Grant.
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks (clearly labeled algorithm sections or code-like formatted procedures).
Open Source Code Yes The code and demo videos are available on our website https://sites.google.com/view/pose-assistedcollaboration.
Open Datasets Yes Specifically, we choose pictures from a texture dataset (Kylberg 2011) and place them on the surface of walls, floor, obstacles etc.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning for validation.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions software components and algorithms (e.g., A3C algorithm, Conv-LSTM network, GRU, CNNs, LSTM Network) but does not provide specific version numbers for these or other software dependencies.
Experiment Setup Yes The action space is discrete and contains eleven candidate actions (turn left, turn right, turn up, turn down, turn topleft, turn top-right, turn bottom-left, turn bottom-right, zoom in, zoom out and keep still). We take a two-phase training strategy for learning. Specifically, we choose pictures from a texture dataset (Kylberg 2011) and place them on the surface of walls, floor, obstacles etc. And we apply the A3C algorithm to update the network architecture of the Pose-Assisted Multi-Camera Collaboration System.