Code for paper: Enhancing Classroom Behavior Recognition with Lightweight Multi-Scale Feature Fusion
conda create -n yolov11 python=3.8.16
conda activate yolov11
pip install torch==2.2.2+cu121
The datasets we used are as follows:
The source dataset was downloaded from the dataset github page https://github.com/Whiffe/SCB-dataset (Accessed on 1th Nov 2024).
Due to database requirements in this paper, we provide forever restricted links access to the reprocessed dataset :
Thanks to the authors for providing the open access STBD-08 and SCB-dataset3 dataset.
You can modify the configuration of the parameters in the XXX.yaml for different dataset.
You can run the bash script as below :
python train.py
Models and results will be saved at folder: 'runs/dataset_name/'.
The results of the trained model can be downloaded directly from this URL:FRNet/runs
Please cite the following paper if you use this repository in your reseach.
@Article{cmc.2025.066343,
AUTHOR = {Chuanchuan Wang, Ahmad Sufril Azlan Mohamed, Xiao Yang , Hao Zhang , Xiang Li, Mohd Halim Bin Mohd Noor},
TITLE = {Enhancing Classroom Behavior Recognition with Lightweight Multi-Scale Feature Fusion},
JOURNAL = {Computers, Materials \& Continua},
VOLUME = {85},
YEAR = {2025},
NUMBER = {1},
PAGES = {855--874},
URL = {http://www.techscience.com/cmc/v85n1/63535},
ISSN = {1546-2226},
DOI = {10.32604/cmc.2025.066343}
}