Multi-class Fish Classification

👩‍💻 Source Code

Description

Convolutional Neural Network (CNN) was used to classify 12 distinct fish species, and core data augmentation significantly improved model performance.

Datasets

Local Fish: 269 images

Catla = 18, Cyprinus carpio = 50, Grass carp = 11, Mori = 70, Rohu = 73, Silver = 47

Shah, Syed Zakir Hussain ; Rauf, Hafiz Tayyab; lali, IkramUllah; Bukhari, Syed Ahmad Chan;  Khalid, Malik Shahzaib; Farooq, Muhammad ; Fatima, Mahroze (2019), “Fish-Pak: Fish Species Dataset from Pakistan for Visual Features Based Classification”, Mendeley Data, V3, doi: 10.17632/n3ydw29sbz.3.

Coastal Fish: 280 images

Black Sea Sprat = 50, Gilt Head Bream = 50, Horse Mackerel = 50, Red Sea Bream = 50, Sea Bass = 50, Trout = 50

O.Ulucan, D.Karakaya, and M.Turkan.(2020) A large-scale dataset for fish segmentation and classification.
In Conf. Innovations Intell. Syst. Appli. (ASYU)

Image Augmentation:

Angle [range(15,180,15)], Horizontal, and Vertical

Augmented 1 image to 24 images

Total dataset = 13176 images

More can be found in the repository.