Skip to content

A PyTorch Implementation of Feature Selective Anchor-Free Module for Single-Shot Object Detection (CVPR'19)

License

Notifications You must be signed in to change notification settings

hdjang/Feature-Selective-Anchor-Free-Module-for-Single-Shot-Object-Detection

Repository files navigation

Feature Selective Anchor-Free Module for Single-Shot Object Detection. CVPR, 2019. (in PyTorch)

Description

This repository reproduces "Zhu et al. Feature Selective Anchor-Free Module for Single-Shot Object Detection. CVPR, 2019." (FSAF) PDF in PyTorch. The implementation is based on MMDetection framework. All the codes for the FSAF model follow the original paper.

Get Started

To use this repo, please follow README.md of MMDetection.

Train/Eval

Train

  • To train baseline (i.e., RetinaNet)
./tools/dist_train_retinanet_r50_400_050x.sh
  • To train FSAF (w/o anchor-based (AB))
./tools/dist_train_fsaf_r50_400_050x.sh

Eval

For evaluation, pretrained model-weights should be located at "./models/here".

  • To evaluate baseline (i.e., RetinaNet)
./tools/eval_retinanet_r50_400_050x.sh
  • To evaluate FSAF (w/o anchor-based (AB))
./tools/eval_fsaf_r50_400_050x.sh

Benchmark

Below is benchmark results. All models are trained with an image-size of 400 and reduced LR-schedule for efficient experiments. Reproduced results show a similar aspect to the original paper (Table 1,2), demonstrating sanity of the implementation.

model backbone img-size LR-schd box AP box AP_50 box AP_75 download
RetinaNet R-50-FPN 400 0.50x 26.0 43.4 27.1 model
FSAF (w/o AB) R-50-FPN 400 0.50x 26.2 44.7 26.5 model

TODO

  • Code reorganization is needed to be consistent with the style of MMDetection framework (current code is only written for fast prototyping).

Contact

About

A PyTorch Implementation of Feature Selective Anchor-Free Module for Single-Shot Object Detection (CVPR'19)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published