DePondFi'24 focuses on detecting fish key points from real-time underwater images to aid intelligent aquaculture systems in fish identification and biomass estimation. This challenge aims to develop robust computer vision algorithms to address the complexities of underwater environments, such as varying lighting, turbidity, and fish occlusions. The initiative supports Sustainable Development Goal 14 by enhancing aquaculture monitoring efficiency, aiding in disease detection, feeding optimization, and overall aquaculture management.
python predict_keypoints_colab.py --model_path "/content/best_model.pth" --input_folder "/content/drive/MyDrive/Images"