Multiclass Cucumber Leaf Diseases Recognition Using Best Feature Selection

Agriculture is an important research area in the field of visual recognition by computers. Plant diseases affect quality and yields agriculture. Early-stage identification crop disease decreases financial losses positively impacts quality. The manual diseases, which are mostly visible on leaves, a very time-consuming costly process. In this work, we propose new framework for cucumber leaf diseases. proposed based deep learning involves fusion selection best features. feature extraction phase, VGG (Visual Geometry Group) Inception V3 models considered fine-tuned. Both fine-tuned trained using transfer learning. Features extracted later step fused parallel maximum approach. step, features selected Whale Optimization algorithm. best-selected classified supervised algorithms final classification experimental process was conducted privately collected dataset that consists five types achieved accuracy 96.5%. A comparison with recent techniques shows significance method.