Classification of dog skin diseases using deep learning with images captured from multispectral imaging device

Abstract Background Dog-associated infections are related to more than 70 human diseases. Given that the health diagnosis of a dog requires expertise veterinarian, an artificial intelligence model for detecting diseases could significantly reduce time and cost required efficiently maintain animal health. Objective We collected normal multispectral images develop classification each three skin (bacterial dermatosis, fungal infection, hypersensitivity allergic dermatosis). The single models (normal image- image-based) consensus were developed used four CNN architecture (InceptionNet, ResNet, DenseNet, MobileNet) select well-performed model. Results For models, such as or image-based model, best accuracies Matthew’s correlation coefficients (MCCs) validation data set 0.80 0.64 bacterial 0.70 0.36 0.82 0.47 dermatosis. MCCs 0.89 0.76 dermatosis set, 0.87 0.63 infection respectively, which supported disease balanced well-performed. Conclusions dogs by combining with images, respectively. Since be determine areas suspected lesion additionally help confirming redness area, achieved higher prediction accuracy performance between sensitivity specificity.