5 May 2017 Thermographic image analysis for classification of ACL rupture disease, bone cancer, and feline hyperthyroid, with Gabor filters
Author Affiliations +
Thermography and pattern classification techniques are used to classify three different pathologies in veterinary images. Thermographic images of both normal and diseased animals were provided by the Long Island Veterinary Specialists (LIVS). The three pathologies are ACL rupture disease, bone cancer, and feline hyperthyroid. The diagnosis of these diseases usually involves radiology and laboratory tests while the method that we propose uses thermographic images and image analysis techniques and is intended for use as a prescreening tool. Images in each category of pathologies are first filtered by Gabor filters and then various features are extracted and used for classification into normal and abnormal classes. Gabor filters are linear filters that can be characterized by the two parameters wavelength λ and orientation θ. With two different wavelength and five different orientations, a total of ten different filters were studied. Different combinations of camera views, filters, feature vectors, normalization methods, and classification methods, produce different tests that were examined and the sensitivity, specificity and success rate for each test were produced. Using the Gabor features alone, sensitivity, specificity, and overall success rates of 85% for each of the pathologies was achieved.
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Mehrdad Alvandipour, Mehrdad Alvandipour, Scott E. Umbaugh, Scott E. Umbaugh, Deependra K. Mishra, Deependra K. Mishra, Rohini Dahal, Rohini Dahal, Norsang Lama, Norsang Lama, Dominic J. Marino, Dominic J. Marino, Joseph Sackman, Joseph Sackman, } "Thermographic image analysis for classification of ACL rupture disease, bone cancer, and feline hyperthyroid, with Gabor filters", Proc. SPIE 10214, Thermosense: Thermal Infrared Applications XXXIX, 102140B (5 May 2017); doi: 10.1117/12.2261753; https://doi.org/10.1117/12.2261753

Back to Top