Morphology-based identification and classification of Pediastrum through AlexNet Convolution Neural Network. Issue 1 (April 2021)

• Record Type: Journal Article
• Authors: Gaur, Ashish; Yadav, D.P.; Pant, Gaurav
• Is Part Of: IOP conference series. Volume 1116:Issue 1(2021)
• Publication Date: 2021-04
• Subjects: Convolution Neural Network (CNN) -- AlexNet -- Image processing -- Coenobia -- Pediastrum -- Deep learning; Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/1116/1/012197
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: In general solutions and outcomes to an issue in image processing, a ton of trail and examination with an enormous assortment of test photographs is incorporated. Pediastrum, which comprises of pigments like carotene, xanthophyll, chlorophylls, and other nourishing elements like complex sugars, major and minor minerals, and protein, enzymes, fiber, is a minute coccal shaped colony forming green algae. In this research, an attempt was made to solve the problem of identifying and classifying different species of Pediastrum with the aid of CNN based deep learning model. The informational collection comprising of 12, 000 algal pictures was being utilized by, AlexNet CNN (Convolution Neural Network) model for the training and validating purposes. The features like the presence of sporopollenin in the cell wall, the structures and the functions of the cells, and the structural properties of coenobia make the basis for automated classification by AlexNet deep learning model. The efficacy of the proposed approach is demonstrated by an experimental outcome of 99.54 percent classification accuracy with precision and Fl-score more than 0.98.