Project College | Informatics Master Degree | Institute of Technology Bandung - http://www.itb.ac.id
Dataset: 390 images
step 1: Color Quantization with KMeans clustering - Distance?
step 2: use ColorDescriptor with HSV region-histogram based
Color
histogram: feature
region-1 >> 288 --> get 50 center
region-2 >> 288 --> get 50 center
region-3 >> 288 --> get 50 center
region-4 >> 288 --> get 50 center
region-5 >> 288 --> get 50 center
Features Center Total: 250 feature-center
390 image x 250 feature-center = 97,500 features-vector (Dimension: 390 rows x 250 cols)
Color Dominant
Shape (CEDD - Lire)
Maker
Year