This is basic content Based Image Retrieval project
- split the dataset into train and test sets
- here i used 80 : 20 as the ratio of train and test split
- convert each image in HSV formate
- divide image into 5 segments : top-left, top-right, bottom-right, bottom-left, center as ecllipse
- histogram bins : here 8, 12, 3 bin sizes
- for 1 segment of the image 8*12*3 features
- for 5 segments 5*8*12*3 == 1,440 -- > this is the feature vector size for the single image
- Divide the image into 5 segments
- Finding feature vector by using bins
- Storing the feature vectors in a bag
- For test Image finding feature vector by using bins
- Measure the simmilarity of the image with the features present in our bag