Comparative Analysis of Formalin Detection in Fish Using Deep Neural Networks

For the people of Bangladesh, fish is a non-vegetarian and staple food. As Bangladesh is a riverine country, fish is available everywhere in our country. Fish rots easily. For this reason, the fish has to be preserved to fulfill people's nutritional needs. To meet the needs of the people of Ban...

Full description

Saved in:

Bibliographic Details
Published in:	2024 3rd International Conference on Advancement in Electrical and Electronic Engineering (ICAEEE) pp. 1 - 6
Main Authors:	Islam, Rashedul, Islam, Sabrin, Jeba, Samiha Maisha, Dihan, Rakibul Hasan
Format:	Conference Proceeding
Language:	English
Published:	IEEE 25-04-2024
Subjects:	Accuracy Deep Neural Network Digital images EfficientNet Fish Formaldehyde in fish Formalin Fish Pain Predictive models Rivers Sociology VGG16
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	For the people of Bangladesh, fish is a non-vegetarian and staple food. As Bangladesh is a riverine country, fish is available everywhere in our country. Fish rots easily. For this reason, the fish has to be preserved to fulfill people's nutritional needs. To meet the needs of the people of Bangladesh, fish is collected from the rivers and sometimes imported for this vast population. While importing fish or taking it from one place of the country to another, unscrupulous traders mixed formalin to prevent the fish from rotting. Formalin is very harmful to the human body. It creates so many health problems like cancer, miscarriage, tumors, etc. So, it is essential to detect formalin fish and non-formalin fish. There is a need for a more dependable and automated method of formalin detection because current approaches rely on human senses and are prone to inaccuracy. Digital image processing has become a promising technology for fish formalin detection in recent years. In this paper, an intelligent application is proposed. This model can identify formalin and fresh fish based on the images of fish eyes. The capability of digital image processing to offer a quick and non-intrusive examination technique is one of the main advantages of formalin detection. Without damaging sample preparation, it can swiftly and accurately identify the presence of formalin by examining photographs of the fish. This is especially helpful in large-scale operations when it is impractical to inspect each fish manually. Digital image processing algorithms are more accurate and consistent because they are not influenced by human biases or subjective judgments like human senses are. The architecture we proposed is VGG-16, and we compare it with EfficientNet-B3. VGG-16, renowned for its simplicity and effectiveness in various computer vision tasks, achieved a notable accuracy of 94.44%.
DOI:	10.1109/ICAEEE62219.2024.10561763