An Efficient Bag-of-Features for Diseased Plant Identification

Agriculture is the main source of human habitation and its redemption from disease is a primary concern for any economy. For the same, computer vision techniques have been proven to be quite useful. However, the diseased plant identification is still a challenging task due to the disparity in the leaf images. To alleviate the same, this chapter proposes a new bag-of-features-based diseased plant identification method. In the proposed method, the efficient visual words are generated using gray relational analysis-based clustering method, and for image encoding, two-dimensional vector quantization method is used. The proposed method is evaluated on a publicly available leaf images dataset, i.e., PlantVillage. Moreover, the performance is compared against the state-of-the-art classification methods in terms of accuracy, precision, recall, sensitivity, and specificity. Experiments validate that the proposed method is efficient than the compared methods image classification.