🚀 AFENet: Adaptive Frequency Enhancement Network for Remote Sensing Image Semantic Segmentation, IEEE TGRS 2025

---

📌 Introduction

This repository contains the official implementation of our paper:
📄 Adaptive Frequency Enhancement Network for Remote Sensing Image Semantic Segmentation, IEEE TGRS 2025

Feng Gao, Miao Fu, Jingchao Cao, Junyu Dong, Qian Du

AFENet is an advanced semantic segmentation network specifically designed for high-resolution remote sensing image segmentation. By integrating spatial and frequency domain features, AFENet dynamically adapts network parameters to various land cover types while enhancing the interaction between spatial and frequency features, achieving high-precision segmentation results and strong generalizability.

🔍 Key Features

🔍 Key Features:
✅ Adaptive Frequency and Spatial Interaction
✅ Dynamic Feature Modulation for Diverse Land Covers
✅ Enhanced Contextual and Detail Representation
✅ Superior Segmentation Accuracy and Generalizability

---

📂 Dataset

The dataset used in our experiments can be accessed from the following link:

📥 [Download Dataset ](Benchmark on Semantic Labeling)

---

🛠 Installation & Dependencies

Before running the code, make sure you have the following dependencies installed:

conda create -n airs python=3.8
conda activate airs
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt

---

Data Preprocessing

After download the datasets, you still need split them yourself! There are some scripts:

Generate the training set.

python GeoSeg/tools/DATASET_patch_split.py \
--img-dir "Your/Image/Path" \
--mask-dir "Your/Mask/Path" \
--output-img-dir "Your/OutPut/Path" \
--output-mask-dir "Your/OutPut/Path" \
--mode "train" --split-size 1024 --stride 512

tips: stride has different values, vaihingen is 512, while potsdam is 1024.

Generate the testing set.

python GeoSeg/tools/DATASET_patch_split.py \
--img-dir "data/vaihingen/test_images" \
--mask-dir "data/vaihingen/test_masks_eroded" \
--output-img-dir "data/vaihingen/test/images_1024" \
--output-mask-dir "data/vaihingen/test/masks_1024" \
--mode "val" --split-size 1024 --stride 1024 \
--eroded

Generate the masks_1024_rgb (RGB format ground truth labels) for visualization.

python GeoSeg/tools/DATASET_patch_split.py \
--img-dir "data/vaihingen/test_images" \
--mask-dir "data/vaihingen/test_masks_eroded" \
--output-img-dir "data/vaihingen/test/images_1024" \
--output-mask-dir "data/vaihingen/test/masks_1024" \
--mode "val" --split-size 1024 --stride 1024 \
--gt

🏋️‍♂️ Usage: Training AFENet

To train AFENet on the Vaihingen/Potsdam dataset, use the following command:

python train.py -c config/vaihingen/afenet.py --max_epoch 100 --lr 6e-4 --batch_size 8

🔧 Training Arguments:

• --epoch: Number of training epochs
• --lr: Learning rate
• --batch_size: Batch size

---

Usage: Testing AFENet

python DATASET_test.py -c config/DATASET/afenet.py -o Your/Output/Path --rgb

📬 Contact

🔥 We hope AFFNet is helpful for your work. Thanks a lot for your attention.🔥

If you have any questions, feel free to contact us via Email:
📧 gaofeng@ouc.edu.cn
📧 fumiao@stu.ouc.edu.cn