Progressive quality estimation of oyster mushrooms using neural network–based image analysis
Tanmay Sarkar, Alok Mukherjee, Kingshuk Chatterjee, Slim Smaoui, Siddhartha Pati, Mohammad Ali Shariati
Abstract
We have developed an artificial intelligence–based quality prediction model for oyster mushroom samples in this work. The proposed model tends to predict the progressively deteriorating quality of the samples in terms of predicted Hedonic number, which is adjudged as one of the most reliable scales of raw fruit quality assessment parameters. The present scheme attempts to continuously assess the quality of mushrooms by judging the extent of deterioration of the sample images; instead of discrete classification asserting only the edibility or non-edibility of the samples. Thus, the extent of the freshness of any test sample could also be approximated using the predicted Hedonic number from the model. The proposed scheme uses an artificial neural network to develop the estimator. The simplicity of analysis of the scheme and high accuracy of prediction of freshness allow for basic screening of the samples without requiring a panel of experts to judge the same, which is a difficult task, especially under this pandemic circumstance. Besides, implementing the proposed algorithm in designing possible mobile-based application software would widen its applicability in a practical scenario.