brandon-hastings / lumeleon Goto Github PK

Implement GrabCut algorithm preceded by kMeans clustering

Output from bounding box selection is csv with image file path used and bounding box coordinates.

#3 needs to be completed before moving this feature to testing stage.

Bounding box of standard is used to create intensity matched image, bounding box of animal is used on intensity matched image to segment from background

• return bounding box file to set location in project folder and set location in grab cut function call

Create multi-class bounding box regression that can be run within google Colab

• implement 20/80 train test split of images in project
• modify image homography labeling toolbox for use with bounding box labelling