International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 185 - Issue 36 |
Published: Oct 2023 |
Authors: Md. Rashedul Islam Mamun, Umma Saima Rahman, Tahmina Akter, Muhammad Anwarul Azim |
![]() |
Md. Rashedul Islam Mamun, Umma Saima Rahman, Tahmina Akter, Muhammad Anwarul Azim . Fish Disease Detection using Deep Learning and Machine Learning. International Journal of Computer Applications. 185, 36 (Oct 2023), 1-9. DOI=10.5120/ijca2023923079
@article{ 10.5120/ijca2023923079, author = { Md. Rashedul Islam Mamun,Umma Saima Rahman,Tahmina Akter,Muhammad Anwarul Azim }, title = { Fish Disease Detection using Deep Learning and Machine Learning }, journal = { International Journal of Computer Applications }, year = { 2023 }, volume = { 185 }, number = { 36 }, pages = { 1-9 }, doi = { 10.5120/ijca2023923079 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2023 %A Md. Rashedul Islam Mamun %A Umma Saima Rahman %A Tahmina Akter %A Muhammad Anwarul Azim %T Fish Disease Detection using Deep Learning and Machine Learning%T %J International Journal of Computer Applications %V 185 %N 36 %P 1-9 %R 10.5120/ijca2023923079 %I Foundation of Computer Science (FCS), NY, USA
Fish farming is the practice of rearing fish in cages for human consumption. It is the area of animal food production that is expanding most rapidly. It thus is vital for higher fish production. However, it is deteriorating as a result of several diseases. Numerous ailments and conditions known to harm fish have documented causes, including concern, overcrowding, poor water quality, and failure to quarantine any recently arrived or ill fish to prevent disease transmission. Untrained farmers have a difficult time spotting fish disease. This issue can be resolved with low-cost fish disease detection equipment. Since we're going in the era of data science, we want to compare which deep learning or machine learning method is better suited for this area of study. To build an accurate model, we gather a total of 1382 images for the four classes of White Spot, Black Spot, Red Spot, and Fresh Fish. When it comes to image classification, deep learning outperforms machine learning. In this study, we used a segmentation technique to locate the afflicted area. With the aid of performance evaluation matrices, nine popular classification algorithms as well as two ensemble methods are also used to measure performance. The highest accuracy, 99.64%, is achieved by the VGG16 and VGG19 ensemble models, while ResNet-50, a pre-trained model, achieved 99.28% accuracy, outperforming Random Forest's 90.25% accuracy.