UFPMP-Det:Toward Accurate and Efficient Object Detection on Drone Imagery

Authors: Yecheng Huang, Jiaxin Chen, Di Huang1026-1033

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments are carried out on the widely used Vis Drone and UAVDT datasets, and UFPMP-Det reports new state-of-the-art scores at a much higher speed, highlighting its advantages.
Researcher Affiliation Academia 1 State Key Laboratory of Software Development Environment, Beihang University, Beijing, China 2 School of Computer Science and Engineering, Beihang University, Beijing, China
Pseudocode Yes Algorithm 1: Foreground Region Generation
Open Source Code Yes The code is available at https://github.com/PuAnysh/UFPMP-Det.
Open Datasets Yes UFPMP-Det is evaluated on the widely-used Vis Drone (Zhu et al. 2018) and UAVDT (Du et al. 2018) benchmarks and extensive experiments are carried out.
Dataset Splits Yes Vis Drone consists of 10,209 high resolution images (2000 × 1500) with 10 object categories... 6,471 images are used for training, 548 for validation and 3,190 for test.
Hardware Specification Yes All the experiments are conducted on one GTX 1080TI GPU.
Software Dependencies No The paper mentions the use of the "MMDetection toolbox" and "GFL" but does not provide specific version numbers for these or any other software dependencies.
Experiment Setup Yes The momentum and weight decay are fixed as 0.9 and 0.0001, respectively. The initial learning rate is set at 0.01 with a linear warm-up, which decreases by the factor of 10 after 40 and 55 epochs.