README.md

**STF: Spatio-Temporal Fusion Module for Improving Video Object Detection **

OpenMMLab website ^HOT      OpenMMLab platform ^{TRY IT OUT}

 

 

📘Documentation | 🛠️Installation |

English | 简体中文

Introduction

STF module is the Fusion Module for Improving Video Object Detection and it is a part of the OpenMMLab project.

The main branch works with PyTorch 1.8+.

Installation

Please refer to Installation for installation instructions.

Getting Started

Please see Overview for the general introduction of MMDetection. You need to follow the steps defined in the docs.

To train the model, use the following command:

python tools/train configs/centernet/hrnet.py

For detailed user guides and advanced guides, please refer to our documentation:

• User Guides

  • Train & Test
    • Learn about Configs
    • Inference with existing models
    • Dataset Prepare
    • Test existing models on standard datasets
    • Train predefined models on standard datasets
    • Train with customized datasets
    • Train with customized models and standard datasets
    • Finetuning Models
    • Test Results Submission
    • Weight initialization
    • Use a single-stage detector as RPN
    • Semi-supervised Object Detection
  • Useful Tools

Datasets

In this repo, we used the following datasets:

1. KITTI MOTS Dataset
2. Cityscapes Dataset
3. UAVDT Dataset

Results

These are the results on Cityscape, KITTI (MOT) and UAVDT, respectively.

 

 

The graph for training loss wrt to the Cityscapes dataset is as below:

 

 

FAQ

Please refer to FAQ for frequently asked questions.

Acknowledgement

We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC),[funding reference number RGPIN-2020-04633].

Citation

If you use this module in your research, please cite this project. (https://arxiv.org/abs/2402.10752)

License

This project is released under the Apache 2.0 license.