- Image Classifier for Cat Breed
- End-Term Project for AI course @ Inha University
- Data Collecting
- Data Augmentation
- Model Architecting
- Integrating Model (Web Service)
- https://www.kaggle.com/ma7555/cat-breeds-dataset
- 70,988 Images (67 Cat Breeds)
- https://www.robots.ox.ac.uk/~vgg/data/pets/
- 2,403 Images (12 Cat Breeds)
- Abyssinian
- American Bobtail
- American Curl
- American Shorthair
- Applehead Siamese
- Balinese
- Bengal
- Birman
- Bombay
- British Shorthair
- Burmese
- Calico
- Cornish Rex
- Devon Rex
- Dilute Calico
- Dilute Tortoiseshell
- Egyptian Mau
- Exotic Shorthair
- Extra-Toes Cat - Hemingway Polydactyl
- Havana
- Himalayan
- Japanese Bobtail
- Maine Coon
- Manx
- Munchkin
- Nebelung
- Norwegian Forest Cat
- Ocicat
- Oriental Short Hair
- Oriental Tabby
- Persian
- Pixiebob
- Ragamuffin
- Ragdoll
- Russian Blue
- Scottish Fold
- Siamese
- Siberian
- Snowshoe
- Sphynx - Hairless Cat
- Tabby
- Tiger
- Tonkinese
- Torbie
- Tortoiseshell
- Turkish Angora
- Turkish Van
- Tuxedo
- Due to small training images ( < 100 )
- York Chocolate
- Chinchilla
- Canadian Hairless
- Burmilla
- LaPerm
- Cymric
- American Wirehair
- Singapura
- Chausie
- Javanese
- Somali
- Oriental Long Hair
- Korat
- Selkirik Res
- Chartreux
- Silver
- Due to difficulty of distinguishing (e.g. Mixed Breeds)
- Domestic Short / Medeium / Long Hair
- 64 Cases (2 * 4 * 2 * 4)
- Blur (Filter Size = 3, 5)
- Noise (Gaussian Noise, STD. Deviation = 5, 10 ,15, 20)
- Translation (Coordinate Translation (0, 0) -> (N, N), N = 5, 10)
- Rotation (Rotation by- 10, -5, 5, 10 Degree)