I. Introduction
Rice Crop plays a prominent role in the Indian agricultural economy. In fact, rice is cultivated in an area of 72.06 million hectares. Rice is a nutritious staple food that offers instant energy, which is rich in carbohydrates [1]. In the production of rice, diseases are often considered one of the critical limiting factors. Rice crops are severely affected by diseases, which leads to a significant decline in the agricultural economy. For Instance, Bacterial Leaf Blight (RBL), Rice Blast (RB), Rice Tungro (RT), and Brown Spot (RS) are the most common diseases which cause significant loss in rice irrigation and cause lesions on rice leaves. Currently, there are no effective methods for preventing rice plant diseases at an early stage. Thus, developing an accurate computer-aided system for the detection and classification of plant diseases is required to diagnose and prevent the spread of diseases, thereby increasing rice production and quality. Pattern recognition in machine learning is mainly comprised of four straightforward approaches, 1. Image Acquisition, 2. Image/Data Preprocessing, 3. Feature Extraction, and 4. Classification. Raw patterns from digital images are acquired during image aquation; noise removal, geometric corrections, illumination corrections, and color space conversions are handled during preprocessing; attributes extraction for pattern classification [2] is handled during feature extraction; and finally, classification classifies given input data to one of the predefined classes based on probability. Khan, et al. [3], concentrated on recognition and classification based on correlation coefficients and depth features (CCDF) of several fetal diseases. Two intensely trained models VGG 16, Caffe AlexNet, were used to detect the signs of selected diseases in apple and banana with fruit spots. A genetic algorithm is used for FS, and M-SVM is used for classification. Accuracy can be improved by using feature extraction techniques.