Python-based OpenCV program for detecting leaves and creating segmentation masks based on images in the Komatsuna dataset.
Set up the dataset from the above link, install the pip requirements and run python3 PlantDetector.py
.
Key | Function |
---|---|
m | next image |
n | prev image |
s | save selected image in the selected mode |
z | save all images in selected mode |
esc | exit the program |
d | dynamically calculate dice |
f | show dice data based on saved images |
1-5 | select the respective mode |
Mode | Window "Altered" | Window "Original" |
---|---|---|
0 | original (fallback) | original |
1 | HSV filter range | original |
2 | bare watershed masks | labels |
3 | watershed masks w/ bg | original |
4 | sequential bg sub | original |
5 | seq bg sub w/ watersh | original |
View the project report below or on Overleaf.