Dried fish dataset for Indian seafood: A machine learning application

Abstract

Dryingfish is a simple and economical way to process the catch. It creates a profitable business for coastal communities by providing a market for their catches, even during periods of abundance. It's a traditional method to preserve fish, especially valuable in regions where fresh fish isn't readily available or affordable throughout the year. This dataset provides a rich resource of 8290 images specifically designed for machine learning applications. It focuses on the five most popular types of dried seafood in India: prawns (shrimp), small anchovies (tingali), golden anchovies (mandeli), mackerel (bangada), and Bombay duck (bombil). To ensure high-quality data for machine learning applications for Identification and classification of different dried fish varieties, the dataset features a diverse set of images in singles and in bulk for each category. The dataset utilizes standardized lighting, background, and object pose for optimal machine learning performance. This rich dataset empowers researchers and data scientists to leverage machine learning for various applications in the Indian dried fish industry.Overall, the Dried Fish Dataset for Indian Seafood aims to leverage machine learning to improve the standardization, quality control, safety, and efficiency of the Indian dried fish industry.

Keywords: Dried fish classification; Dried fish dataset; Dried fish detection; Machine learning.

Fig. 1

Organization of dataset.

Fig. 2

Image capturing process.

Fig. 3

Data acquisition steps.