This repo contains sample algorithms that work with eCAL and the Mediapipe framework. It demonstrates how to perform tasks based on subscribing to eCAL data (in form of images) and publishing the results.
- object-detection A sample that shows how to subscribe to images, detect objects on them, and publish the detected bounding boxes.
- gesture-recognition A sample that shows how to subscribe to images, recognizes hands and their gestures and publishes the results and hand connections.
- image-segmentation A sample that shows how to divide images in different regions. It recognizes people and fills an image with pre-defined colours in the divided regions (person, background).