Decoupling Classifier for Boosting Few-shot Object Detection and Instance Segmentation

Authors: Bin-Bin Gao, Xiaochen Chen, Zhongyi Huang, Congchong Nie, Jun Liu, Jinxiang Lai, GUANNAN JIANG, Xi Wang, Chengjie Wang

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

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we empirically evaluate the proposed method for FSOD/g FSOD and FSIS/g FSIS tasks and demonstrate its effectiveness by comparison with state-of-the-art methods.
Researcher Affiliation Industry Bin-Bin Gao1 Xiaochen Chen1 Zhongyi Huang1 Congchong Nie1 Jun Liu1 Jinxiang Lai1 Guannan Jiang2 Xi Wang2 Chengjie Wang1 1Tencent You Tu Lab 2CATL
Pseudocode No The paper describes its methodology through mathematical equations but does not include any pseudocode or algorithm blocks.
Open Source Code Yes 1https://csgaobb.github.io/Projects/DCFS. Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] See Section 4 and https://csgaobb.github.io/Projects/DCFS.
Open Datasets Yes We follow the previous works and evaluate our method on PASCAL VOC [4] and MS-COCO [21] datasets.
Dataset Splits Yes For a fair comparison, we use the same data splits given in [33, 28].
Hardware Specification Yes The experiments are conducted with Detectron2 [36] on NVIDIA GPU V100 on CUDA 11.0.
Software Dependencies Yes The experiments are conducted with Detectron2 [36] on NVIDIA GPU V100 on CUDA 11.0.
Experiment Setup Yes The SGD is used to optimize our network end-to-end with a mini-batch size of 16, momentum 0.9, and weight decay 5e 5 on 8 GPUs. The learning rate is set to 0.02 during base training and 0.01 during few-shot fine-tuning.