Introduction: The bilaterally symmetry property in animals can be used to detect pathologies where body parts on both sides can be compared. For any pathological disorder, thermal patterns differ compared to the normal body parts. A software application for veterinary clinics is under development to input two thermograms of body parts on both sides, one normal and the other unknown, and the application compares them on the basis of extracted features and appropriate similarity and difference measures and outputs the likelihood of pathology. Previous research has been used to determine the appropriate image processing, feature extraction and comparison metrics to be used. The comparison metrics used are the vector inner product, Tanimoto, Euclidean, city block, Minkowski and maximum value metric. Also, results from experiments with comparison tests are used to derive a potential threshold values which will separate normal from abnormal images for a specific pathology.
Objectives: The main objective of this research is to build a comparison software tool application by combining the concepts of bilateral symmetrical property in animals and IR thermography that can be for prescreening in veterinary clinics.
Comparison Software Tool Development: The comparison software tool was developed for veterinary clinics as a prescreening tool for pathology detection using the concepts of thermography and bilateral symmetry property in animals. The software tool has a graphical user interface (GUI) that allows ease of use for the clinical technician. The technician inputs images or raw temperature data csv files and compares thermographic images of bilateral body parts. The software extracts features from the images and calculates the difference between the feature vectors with distance and/or similarity metrics. Based upon these metrics, the percentage deviation is calculated which provides the deviation of the unknown (test) image from the known image. The percentage deviation between the thermograms of the same body parts on either side provides an indication regarding the extent and impact of the disease [Poudel; 2015]. The previous research in veterinary thermography [Liu; 2012; Subedi; 2014, Fu; 2014, Poudel; 2015] has been combined with the real world veterinary clinical scenario to develop a software tool that can be helpful for researchers as well as for the clinical technicians in prescreening of pathologies.
Experimental Results and Discussion: Experiments were performed on ACL thermograms to determine a threshold that can separate normal and abnormal ACL images. 18-colored Meditherm images had poor results and could not suggest any threshold value. But results were positive for temperature remapped 256 gray level Meditherm images which suggested the 40% of percentage deviation could produce a separation. The total number of Normal - Normal pairs were greater than total number of Normal – Abnormal pairs below 40% deviation. Similarly, total number of Normal –Abnormal pairs of images were greater than total number of Normal – Normal pairs above 40%. This trend was consistent for Euclidean distance, maximum value distance and Minkowski distance for both texture distances of 6 and 10. The performance in terms of sensitivity and specificity was poor. The best sensitivity of 55% and best specificity of 67% was achieved. This indicates better results for predicting the absence of ACL rupture then actually finding the disease. In this case the software could be used by the clinician in conjunction with other diagnostic methods.
Conclusion: The Experiments, results and analysis show that the comparison software tool can be used in veterinary clinics for the pre-screening of diseases in canines and felines to estimate the extent and impact of the disease based upon the percentage deviation. However, more research is necessary to examine its efficacy for specific pathologies. Note that the software can be used by researchers to compare any two images of any formats. For ACL experimentation, there are indication that a threshold value is possible to separate normal from abnormal and the spectral, texture and spectral features suggested by researches [Subedi; 2014, Liu; 2012, Fu; 2014, Poudel; 2015] are not sufficient to determine that threshold with the given image database.
|Advisor:||Umbaugh, Scott E.|
|Commitee:||LeAnder, Robert, Shang, Ying|
|School:||Southern Illinois University at Edwardsville|
|Department:||Electrical and Computer Engineering|
|School Location:||United States -- Illinois|
|Source:||MAI 56/05M(E), Masters Abstracts International|
|Subjects:||Computer Engineering, Biomedical engineering, Electrical engineering|
|Keywords:||Anterior cruciate ligament rupture, Clinical applications, Computer vision and image processing, Prescreening tools, Thermography, Veterinary medical application|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be