Warts Disease Detection and Classification in Dogs: A Comprehensive Study Integrating Image Processing Techniques and SVM Classification

Abstract

In these papers, the current research focuses on the detection and classification of warts disease in dogs through the application of image processing techniques. These methods aid in identifying and categorizing warts disease specifically in dogs. The study not only quantifies the prevalence of warts in dogs but also categorizes the number of dogs experiencing issues related to this condition. Warts, also referred to as papillomas, represent a common dermatological concern in dogs, drawing attention due to their distinctive appearances and varied clinical manifestations. The research provides a thorough examination of warts, delving into their causes, clinical features, diagnostic methods, treatment options, and their implications in veterinary practice. The underlying cause of warts is primarily linked to viral infections, specifically within the papillomavirus family. These benign growths manifest in various forms, from solitary protrusions to clustered structures resembling cauliflower, with a higher prevalence in younger dogs and those with compromised immune systems. Typically found on mucous membranes, lips, mouth, and occasionally on the skin, warts generally pose minimal health risks, though their presence can lead to discomfort and functional limitations. Accurate diagnosis of canine warts relies on clinical evaluation, often supported by histopathological examination to confirm the viral origin. Treatment options include spontaneous regression, surgical excision, cryotherapy, and in some cases, immunomodulatory therapies. The most suitable approach to wart management depends on factors such as wart location, size, and the overall health status of the animal. The research also involves the analysis of wart samples using image processing techniques within MATLAB. The results showcase original images of areas affected by warts in dogs, followed by segmented output images. The proposed algorithm significantly improves detection accuracy, achieving an enhanced accuracy of 95% in the classification phase using the Minimum Distance Criterion with K-Means Clustering. In the subsequent classification phase, a Support Vector Machine (SVM) classifier is employed, demonstrating a high accuracy of 99% in identifying the presence of warts in dogs. These findings suggest that the proposed algorithm, especially when combined with the SVM classifier, surpasses other methods, substantially improving the accuracy of detecting skin diseases in dogs compared to previously employed classification methods or algorithms.