This is the implementation of Automated Vision-Based Diagnosis of Banana Bacterial Wilt Black sigatoka diseases.
The project was built on python platform with opencv2 and sklearn.
To set up the shapefeatures library, first compile it by running
./build.sh
Utils_shape -> creates shape feature vectors.
Run extract_shape -> to save the files
find_best_feature -> selects the best shape features based on feature importance
Utils_shape_and_color -> creates color and shapefeature vectors.
Run Auc -> output are the roc-graphs
Run Predict.py -> to predict the class of an image. Supply a sample image.
The list of attributes which can be chosen from is as follows:
0: Area
1: Area of min. enclosing rectangle
2: Square of diagonal of min. enclosing rectangle
3: Cityblock perimeter
4: Cityblock complexity (Perimeter/Area)
5: Cityblock simplicity (Area/Perimeter)
6: Cityblock compactness (Perimeter^2/(4*PI*Area))
7: Large perimeter
8: Large compactness (Perimeter^2/(4*PI*Area))
9: Small perimeter
10: Small compactness (Perimeter^2/(4*PI*Area))
11: Moment of Inertia
12: Elongation: (Moment of Inertia) / (area)^2
13: Mean X position
14: Mean Y position
15: Jaggedness: Area*Perimeter^2/(8*PI^2*Inertia)
16: Entropy
17: Lambda-max (Max.child gray level - current gray level)
18: Gray level
Attributes that were considered here are [12,9,13,10,11,1]