In this project we present a approach to do texture image recognition by performing the vlad of all the texture images based on the (CUReT) texture database. The approach consists of following steps:

1. Database prepossessing; https://github.com/chialingwang/ML_Project/tree/master/1_extract_patch/scrip_gen.csh

2. Apply Mini-batch kmeans to find the centroids; https://github.com/chialingwang/ML_Project/tree/master/2_extract_centroids_withscikit/scrip_gen.csh

3. Get Vlad of each image to find out the feature of image; https://github.com/chialingwang/ML_Project/blob/master/3_get_vlad_withscikit_new/scrip_gen.csh

4. Texture classification by using knn provided from scikit-learn. https://github.com/chialingwang/ML_Project/blob/master/4_classification_withscikit/scrip_gen.csh