Providing an user-friendly interface for building computer vision algorithms. By moving and linking "nodes" representing different OpenCV functions, the user will build his algorithm without actually writing lines of code.
I want it to be extensible, so the nodes are packaged into DLLs and loaded dynamically at runtime (spices up the development)
Features currently working :
- Loading nodes from DLLs
- Adding new nodes to a graph and moving them around
- Connecting nodes together by their "pins" (which represent nodes inputs/outputs)
- Running a live simulation (which updates automatically at each action on the graph)
- Displaying inputs and outputs for each node
- Deleting nodes and connections
- Scrolling in the graph canvas to have more space (no zoom in/out for the moment)