Morphi Lab is a project that explores the integration of image segmentation, depth estimation, and image generation techniques to enhance the generation of 2.5D images.
The Morphi Lab repository has the following structure:
morphi_lab/: Contains models and classes for performing segmentation, depth estimation, and image generation.
- models/: Contains various detection and segmentation models.
- segmenter.py: Main segmentation class module.
- depther.py: Main depth estimation class module.
- test_runs/: Contains wrapper code for testing Morphi Lab functionality on images, saved videos, and live videos.
This project utilizes Python poetry for its installation process, following standard protocols.
To run detection and segmentation an appropriate yolo8 model (e.g,. yolov8n-seg
) must be installed in the morphi_labs/models
directory.
Record video from laptop for testing
python test_runs/test_av/record_live_video.py --output_path=${PWD}/test_data/test_input/test_video_segmented.avi
Live cup segmentation
python test_runs/test_segmenter/live_video.py