Categorization of Fruit images using Artificial Bee Colony Algorithm based on GLCM features

The manual recognition of observed items from the records, which takes a long time. We suggest a neural network approach to classify apple defects. Due to the various fruit varieties classification is a challenging task. We suggested a system of classification based on an artificial bee colony algorithm (FSCABC) and a neural feeding network to classify fruit more precisely (FNN). First, fruit photographs were purchased using a digital camera and then, using a break and fuse algorithm, the backdrop of each image was stripped. To catch the fruits, we used a square window, where we downloaded the images in 256 x 256. Second, each fruit image was derived from a color histogram, texture and shape features to make a feature space. Thirdly, a major components method was used to minimize the dimensions of function space. The function is reduced and the weights and preferences are eventually trained on the FNN of the FSCABC algorithm. We have used the K-fold cross-validation approach to improve the capacity of the FNN. Test findings revealed that the SAFCABC-FNN achieved a precision of 89.1 percentage in 1653 fruit imagery in all 18 categories. The rating accuracy was more than the GM-FNN (GA-FNN) with a rating of 84.8 percentages, 87.9 percentages, the PSO-FNN with 85.4 percentages, the ABC-FNN with 88.2 percentages and the supporting kernel vector. Then FSCABC-FNN was regarded as effective in the fruit classification.