Whale Optimized MLP Neural Network And Enhanced Region Growing for Food Product Inspection

Abstract

Nowadaysthere is the requirement for the growth of specific, fast and intention quality assertion to the nature of food and horticultural food items, because it is difficult to sustain and categorize the food products in elevated quality and protected mode for the growing population. Viewed from the prespective, this paper develops a novel method to resolve  difficulties and to categorize the food as a defective or quality product. Further this proposed method contains four sections such as pre-processing, segmentation, feature extraction and classification. Initially, the input database images are pre-processed by means of histogram equalization technique. Then the feature extraction is applied to get rid of the unique characteristics of the food product. After ERG segmentation process is used to isolate several segmented database images. The MLP neural network classification procedure is employed to inspect the quality of the food. The Whale optimization algorithm used here to train the MLP neural network. The proposed MLP-WOA method exhibits accuracy improvement when compared to the existing methods like MLP-GA, MLP-PSO, MLP-FA and MLP-ABC and in addition to this, the error produced is very less in the proposed method compared to all the existing methods. The proposed method is executed in the operational platform of MATLAB and the results are observed by the available methods.