European Journal of Molecular & Clinical Medicine ISSN 2515-8260 Volume 7, Issue 4, 2020 1355 Plant Disease Identifer Using K-Means and GLSM in Convolution Neural Network S.P. Vijaya Vardan Reddy 1 , T. Suresh 2 , K. Naresh Kumar Thapa 3 , V. Ramkumar 4 , S. Mahabhoob Basha 5 , Deepika.Y 6 1, 2,3,4,5 R.M.K. Engineering College, Kavaraipettai, Chennai, Tamilnadu, India. 6 TATA Consultancy Services, Chennai, Tamil Nadu,India. Abstract— Produces from agriculture which feeds the entire population is dependent on proper farming practices. The growth of technology must pay a way for increasing the produce per acre and also help in reducing the onset of frequently affecting plant disease. Timely help in detecting the diseases coupled with solution helps in productivity and quality of the produce. This paper aims to detect the plant leaf disease based on image detection and using machine learning to identify the disease with accuracy and suggest the solution. The product must cater to the needs of urban and rural farmer and also the person with only lay man knowledge of taking photo. This project mainly focuses on leaf disease like Anthracnose, Bacterial Blight, Cercospora, Alternaria Altermata diseases in the Pomegranate, Indian Beech, Tobacco, and Bitter Gourd leaves. This project aims to identify the disease even with lesser region of Interest and predict the leaf diseases using Convolutional Neural Network Algorithm. Keywords—Image Processing, neural network, image segmentation, Convolutional Neural Network Algorithm 1. Introduction Produces from the agriculture fields is plagued by pests and diseases which reduces the product quality and quantity. Even though farmers use enough amount of pesticides during the growth cycle, the problem persists. Timely identification of the diseases and its treatment is the biggest challenge faced by the novice farmers. Expert’s opinion about detection of disease both native and non-native even in the remote areas will benefit the farmers. This project intends to have a strong dataset of various diseases and provide identification of the disease using image processing coupled with machine learning to provide us with basic knowledge to treat the disease. The dataset of the disease that are used in training are Alternaria Alternate Lycopersicon, Anthracnose, Bacterial blight, Cercospora. The dataset is divided into pathogen, fungal disease and bacterial disease. Alternaria Altermata is a pathogen that causes leaf spot in lot of plant species. Anthracnose is a fungal disease that attack leaves of plants during cool and wet weather. Bacterial blight is caused by the bacterial pathogen which causes pale green spots in leaves. This project has focused on lesser number of diseases with a greater number of photos to train the dataset. Going forward, we shall add more and more disease in the dataset. In this project the major focus is on decision making algorithm. The proposed system consists of following steps: image capturing, extracting the deformed or decolored leaf part from image, which is the Image Processing cycle. We convert the image into a digital format and perform enhanced processing and extract valuable information. These digitized images are then feed into training network and the output is