# MASFNet **Repository Path**: beautiful_corridors/MASFNet ## Basic Information - **Project Name**: MASFNet - **Description**: No description available - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2025-07-24 - **Last Updated**: 2025-07-24 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README ### πŸ“– MASFNet: Multiscale Adaptive Sampling Fusion Network for Object Detection in Adverse Weather     ![visitors](https://visitor-badge.laobi.icu/badge?page_id=PolarisFTL.MASFNet)
--- ![network](https://github.com/PolarisFTL/MASFNet/blob/main/figs/network.png) _An overview of the proposed MASFNet. MASFNet consists of four parts: 1) FAENet, 2) Backbone, 3) MSFNet, and 4) DH. Among them, the FAENet utilizes the Laplacian pyramid decomposition to split the input image into two different components, a low-frequency component (LF) and a high-frequency component (HF). The feature information of the input image is then adaptively enhanced through modular processing. Then, the output of FAENet is fed into the backbone for feature extraction. The backbone eventually outputs two different scale feature maps into MSFNet for multi-scale fusion. Finally, the DH detects targets and calculates the loss to optimize the model._ #### πŸ˜Άβ€πŸŒ«οΈ Experiments ![](https://github.com/PolarisFTL/MASFNet/blob/main/figs/mist.png) ![](https://github.com/PolarisFTL/MASFNet/blob/main/figs/mid-foggy.png) ![](https://github.com/PolarisFTL/MASFNet/blob/main/figs/high-foggy.png) ![](https://github.com/PolarisFTL/MASFNet/blob/main/figs/low-light.png) ![](https://github.com/PolarisFTL/MASFNet/blob/main/figs/exdark.png) #### πŸ“’ News #### πŸ”§ Requirements and Installation > - Python 3.6.2 > - PyTorch 1.8.0 > - Cudatoolkit 11.1.1 > - Numpy 1.17.0 > - Opencv-python 4.1.2.30 #### πŸ‘½ Installation ``` # Clone the MASFNet git clone https://github.com/PolarisFTL/MASFNet.git # Install dependent packages cd MASFNet ``` #### πŸš— Datasets | Dataset Name | Total Images | Train Set | Test Set | Google Drive | BaiduYun | | ------------ | ------------ | --------- | -------- | --------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ | | RTTS | 4,322 | 3,889 | 433 | [Link](https://drive.google.com/file/d/1BhU8NnNIQP0mhzB3F-sh7S8xdu-wTO3D/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1TiRYXcDEwnGst5QBZo2twg) | | ExDark | 7,363 | 6,626 | 737 | [Link](https://drive.google.com/file/d/1Q1oHGJys7KsO_n0JBtHLOns2as8-0p1M/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1Fi9AUdB1HPBbktt6-8SKDQ) | | VOC-Rain | 10,653 | 9,482 | 1,171 | [Link](https://drive.google.com/file/d/1I64t88Oc4yHf8J_U9WghKLYC6pmkM_vz/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1q2iq-cDS0vVm0ZYzLVHAPA) | | VOC-Snow | 10,653 | 9,482 | 1,171 | [Link](https://drive.google.com/file/d/1I64t88Oc4yHf8J_U9WghKLYC6pmkM_vz/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1q2iq-cDS0vVm0ZYzLVHAPA) | #### 🎈 Training and Testing Run the following commands for training & testing:\ 🐻 You need to download the pre-training weights and datasets firstly. | Name | Google | BaiduYun | | -------------- | --------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | | VOC07+12+COCO | [yolov4_tiny_weights_voc.pth](https://drive.google.com/file/d/1DGszoaiVAACPGZBHL-8qg8or15of153y/view?usp=drive_link) | [yolov4_tiny_weights_voc.pth (key:1234)](https://pan.baidu.com/s/1sJW8wYbzIprWvFWQotsLFQ) | | COCO-Train2017 | [yolov4_tiny_weights_coco.pth](https://drive.google.com/file/d/1Y2M-nUEL_cHnQeLzgJO_sFBKTegUUAPq/view?usp=drive_link) | [yolov4_tiny_weights_coco.pth (key:1234)](https://pan.baidu.com/s/10Oo5EwQuh2WHwjRt4MBQ6w) | ```python # train MASFNet for RTTS dataset 1python tools/voc_annotations.py # VOCdevkit_path='the path of RTTS dataset', data_name='rtts' modify the config.py # data_name='rtts' python train.py # during training, the result will be saved in the logs-rtts ``` ```python # eval MASFNet for RTTS dataset python tools/get_map.py # data_name='rtts, # vocdevkit_path='the path of RTTS datase' # model_path = 'los-rtts/best_epoch_weights.pth' python tools/fps.py # compute the speed of model python tools/predict.py # try to predict the image in adverse weather ``` The steps are the same if training other datasets. #### πŸ”₯Model Performance | Method | Dataset | Params | FLOPs | FPS | mAP (%) | Google Drive | BaiduYun | | ------------ | -------- | ------ | ----- | --- | ------- | --------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ | | MASFNet-Fog | RTTS | 6.0M | 10.7G | 152 | 73.68 | [Link](https://drive.google.com/file/d/13tMYePzn9yRMNpl7j6gg6057gAZTzUeZ/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1yIZHZBx9yjmm4bRgCnUDmg) | | MASFNet-Dark | ExDark | 6.0M | 10.7G | 125 | 63.80 | [Link](https://drive.google.com/file/d/1cr4mUwMeppQaGVf9tZLKmDZarbDuyuRA/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1ZZrQYtvgC91yDglnOMdKMg) | | MASFNet-Rain | VOC-Rain | 6.0M | 10.7G | 213 | 60.13 | [Link](https://drive.google.com/file/d/1xerYUIv30YTKdhMcHclgmFDHTJNaoxFm/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1EF2BAAMx04_9RJCqqfOlvQ) | | MASFNet-Snow | VOC-Snow | 6.0M | 10.7G | 214 | 59.52 | [Link](https://drive.google.com/file/d/1x3N6OOSOsP4IE8leVx-02W2qQfBjSJcD/view?usp=drive_link) | [Link (key:1234)](https://pan.baidu.com/s/1Ui0GpmqAwfi7A-F6k0ML3Q) | #### πŸ”—Citation If this work is helpful for your research, please consider citing the following BibTeX entry. ``` @ARTICLE{10955257, author={Liu, Zhenbing and Fang, Tianle and Lu, Haoxiang and Zhang, Weidong and Lan, Rushi}, journal={IEEE Transactions on Geoscience and Remote Sensing}, title={MASFNet: Multiscale Adaptive Sampling Fusion Network for Object Detection in Adverse Weather}, year={2025}, volume={63}, number={}, pages={1-15} } ``` #### πŸ“¨ Contact If you have any questions, please feel free to reach me out at polarisftl123@gmail.com #### 🌻 Acknowledgement This code is based on [YOLOv4-Tiny](https://github.com/bubbliiiing/yolov4-tiny-pytorch.git) & [DENet](https://github.com/NIvykk/DENet.git). Thanks for the awesome work.