Lumpy Skin Disease Detection in Cows Using Resnet-50

Artificial intelligence is one of the emerging domains where a huge number of applications are being developed to replace human experts and introduce intelligent systems to solve many real-world problems. Lumpy skin disease (LSD) is the most widely spreading disease that affects cows and causes economic losses to the agricultural industry. Due to a lack of veterinary experts, it is becoming difficult to diagnose the disease in its early stages, which leads to the deaths of cows, mostly in Ethiopia, which suffers from a lack of livestock due to the huge spread of the LSD virus in cattle. The proposed machine learning model will classify cow images as lumpy or non-lumps. Initially, an image dataset of cows 80 percent is utilized for training and 20 percent for testing. These images are further processed to reduce noise and intensity using Gaussian filtering and histogram equalization. Then the model will extract features from preprocessed images using a convolutional neural network because it is the one profound for image processing and hence used it. After feature extraction, the Resnet-50 algorithm is trained with the features extracted, and Resnet-50 will classify images into lumpy and non-lumps. The existing models forecast the LSD virus using metrological attributes for predicting the occurrence of LSD. Quantitatively, an overall accuracy of 87 percent using random forest. Resnet-50 Model Classifies the LSD disease in cows using the cow images dataset. Quantitatively, an overall accuracy of 93 percent using the Resnet 50 algorithm.