Human hands detection and segmentation

Project for the course Computer Vision, UniPD 2022.

Usage

Model

• Download our custom Dataset (link) and upload it saved as custom_dataset.zip file on your Google Drive, in the main folder MyDrive.

• Upload also the custom_data.yaml file in Google Drive as done before. It will be useful for YOLOv5 configuration.

• Open the Jupyter Notebook handdetection.ipynb on Google Colaboratory and connect your drive in order to run the code properly.

• At the end of the notebook, run the cell to download the model in .onnx format as best.onnx and place it in the directory Model/. It will be used for the C++ code.

References

YOLOv5: object detection architecture and pretrained models use in this project for fine-tuning

Code:

• Python Hand Detection/Segmentation: https://github.com/BlueDi/Hand-Detection

• Tensorflow: https://medium.com/nerd-for-tech/building-an-object-detector-in-tensorflow-using-bounding-box-regression-2bc13992973f

Paper:

• Hand recognition structure: https://www.researchgate.net/publication/282956557_Real_time_finger_tracking_and_contour_detection_for_gesture_recognition_using_OpenCV

Datasets:

• Hand Dataset: https://www.robots.ox.ac.uk/~vgg/data/hands/

• EgoHands: http://vision.soic.indiana.edu/projects/egohands/

• Kaggle Dataset: https://www.kaggle.com/datasets/armannikkhah/hand-dataset

• SynthHands Dataset: https://handtracker.mpi-inf.mpg.de/projects/OccludedHands/SynthHands.htm