Model for Tomato Plant Diseases based on Space Invariant Artificial Neural Networks (SIANN)

Abstract

Crop disease detection is crucial for crop output and agricultural production. Deep learning algorithms have emerged as a prominent research topic for solving the detection of agricultural illnesses.   This research suggested a deep convolutional neural network (CNN) with an attention mechanism system that can better adapt to the identifies of a variety of tomato leaf ailments. The large part of the network structure is made up of residual blocks and attention extraction modules. The model is capable of accurately extracting intricate characteristics of a variety of diseases. In the sphere of agriculture, the automatic detection and diagnosis of tomato plant illnesses is critical.  Prior to computer vision, Deep Convolutional Neural Networks (CNN) or Space Invariant Artificial Neural Networks (SIANN) made substantial improvements in a number of fields, including classification, object identification, and segmentation, with accuracy that was superior to human perception. SIANN faces numerous obstacles, such as computing burden and energy, in order to be used in mobile phones and embedded systems, in addition to its achievements in computer vision tasks.