Deformable Part Region Learning for Object Detection

Authors: Seung-Hwan Bae95-103

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

Reproducibility Variable Result LLM Response
Research Type Experimental Without bells and whistles, our implementation of a Cascade deformable part region detector achieves better detection and segmentation m APs on COCO and VOC datasets, compared to the recent cascade and other state-of-the-art detectors.
Researcher Affiliation Academia Seung-Hwan Bae Vision and Learning Laboratory, Inha University, Korea shbae@inha.ac.kr
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not include an unambiguous statement about releasing open-source code for the described methodology or a direct link to a code repository.
Open Datasets Yes Our D-PRD and Cascade D-PRD are evaluated on MSCOCO17 (Lin et al. 2014) and PASCAL VOC07/12 (Everingham et al. 2015) datasets.
Dataset Splits No The paper mentions using COCO and VOC datasets and their trainval/test sets but does not explicitly provide specific percentages, sample counts for training, validation, and test splits, or explicit citations for the exact splits used for reproduction.
Hardware Specification No The paper does not specify any particular hardware used for running the experiments (e.g., specific GPU/CPU models, memory, or cloud instances).
Software Dependencies No The paper mentions "We use the Detectron2." but does not provide specific version numbers for Detectron2 or its underlying software dependencies like Python, PyTorch, or CUDA.
Experiment Setup Yes We use the default learning schedules 1x or 3x ( 12 or 37 COCO epochs) of Detectron2 for all the evaluation below. Also, all other setting parameters for training and testing are same to those of Detectron2. We set the Io U threshold to (0.5, 0.6, 0.7) from the first to last stage.