Veterinary Thermographic Image Analysis
[pic]
Veterinary Thermographic Image Analysis
Project Number 7- 64878
Report Number 4878-9
Scott E Umbaugh, BSE, MSEE, PhD
Patrick Solt, BSEE, MSCS, PhD Candidate
Hari Krishna Akkineni, BSEE, MSEE Candidate
Computer Vision and Image Processing Laboratory
Southern Illinois University Edwardsville
January 22, 2010
Fall 2009
Submitted to:
Dr. Catherine Loughin
Dr. Dominic Marino, Chief of Staff
Long Island Veterinary Specialists
TABLE OF CONTENTS
Executive Summary…………………………………………………... 2
Introduction and Overview…………...………………………………. 3
Materials and Methods……………………………………………….. 3
Images………………………………………………………..… 3
Software……………………………………………………….. 3
Mask Creation Software…….……………………………5
Pattern Classification Software....………………………..5
Texture Features………………………………………….5
Algorithm Test and Analysis Tool ……………………..5
Experimental Results and Discussion………………………………… 6
Preliminary Experiments………………………………….…… 6
Final Experiments………………………………………………6
Results from Experiment Set #1…………………………..……7
Results from Experiment Set #1 PCA Scatterplot……….…… . 8
Results from Experiment Set #2…………………………..…… 9
Results from Experiment Set #2 PCA Scatterplot ..…….…… 10
Conclusions ………………………………………………………… 11
Current aned Future Projects Overview………………………………12
Future Tasks…………………………………………………………. 13
References……………………………………………………………. 14
Executive Summary
Research and development was continued to investigate the application of image analysis and pattern classification to new thermographic images of 120 canines. Specifically, to examine the thermographic images of canines of the breed Cavalier King Charles Spaniel to investigate the Chiari malformation, or COMS, pathology, and to compare to previous preliminary results. Pattern classification algorithms were developed for for: severe, moderate and mild classes of the pathology. The eventual goal for the research is to be able to differentiate normal and abnormal thermographic patterns in canines as a diagnostic and research tool.
Development of application-specific software was continued. Software from Phase I and Phase II of the research was enhanced. The capability of automatically performing leave-one-out cross-validation is being added to the pattern classification GUI. The software used the CVIPtools image libraries. Continued development and enhancement of this software will be quite useful for this research area.
Two sets of experiments were performed for a total for of 4096 experiments. Each experiment consisted of a specific body part, specific classes and specific features. Experiments were performed with all permutations of the ten selected features on each selected body part.
The first set of experiments used the clinical images that were classified as severe and moderate classes of the pathology. Here, the front of head (A1), and top of head (A1D) images were used. With top and front images, 83% successful classification was achieved. These results are not as good as the 100% previously achieved with smaller sets. This indicates that the class domains are more diverse, and the classes more correlated, than was found with the smaller set. Additionally, further experimentation with color normalization is warranted. However, the 83% success still indicates that there is a difference between the moderate and severe classes with the clinical images. The features texture inverse difference and texture energy are the most predictive features.
The second set of experiments used the clinical images that were classified as severe, moderate and mild classes of the pathology. Here, the front of head (A1), top of head (A1D) images were used. With top and front images, 65% successful classification was achieved. These results are inconclusive but indicate that there may be a difference between the mild, moderate and severe classes in the clinical images.
After the first two sets of experiments were performed comparisons of success rates were made between results on the images. Comparing the results we find that the two separate classes in the clinical (moderate and severe) images are better separable than with the three classes (mild, moderate and severe) in the clinical images. In other words, including the mild class confuses the classifier due to overlap in the mild and moderate classes.
Introduction and Overview
Thermographic images of 120 canines were analyzed using feature extraction and pattern classification tools. The purpose of this research is to examine the thermographic images of canines of the breed Cavalier King Charles Spaniel to investigate the Chiari malformation, or COMS, pathology.
Four questions were investigated and related experiments were performed:
1. Can we successfully differentiate between the classes moderate and severe in the clinical images? Here two different experiments were performed:
➢ Use all front of head images (A1)
➢ Use all top of head images (A1D)
2. Can we successfully differentiate between the classes mild, moderate and severe in the clinical images?
➢ Use all front of head images
➢ Use all top of head images
Materials and Methods
Images
The thermal images are from Long Island Veterinary Specialists taken with a Meditherm Med2000 IRIS. The images are in TIF file format as RGB images with 319 columns by 255 rows, 8-bits per pixel per color band. Note that a total of 18 colors are used in these images. The images are thermal images of 120 canines of the breed Cavalier King Charles Spaniel. Previous experimentation used a separate set of 34 canine images. Included in the data set are images of front of head and top of head.
Software
The CVIPtools (Computer Vision and Image Processing Tools) software was used to investigate and analyze the images. The primary tools used include Analysis->Features and Analysis->Pattern Classification. The features considered include histogram, and texture features. The histogram features include mean, standard deviation, skew, energy and entropy; and the texture features include energy, inertia, correlation, inverse difference, and entropy. The pattern classification tools utilized include various normalization methods, distance and similarity measure and three classification methods.
In the course of the image analysis it was determined that the previously developed special purpose software would be further enhanced for this phase of the research. The software is designed to ease the analysis process and enable us to perform a significantly greater number of experiments. The software that creates image masks for the various body parts was enhanced to include inclusion of specific colors in the thermographs. The application specific software previously developed has functionality to ease the mask creation task and also allows the user to load images, classify images, select features and perform pattern classification experiments (see Figure 1). The application software is being enhanced to automatically run experiments with all permutations for the feature set(s) and to perform leave-one-out cross-valation for all the experiments performed.
[pic]
[pic][pic]
Figure 1. Application specific software developed for this project
Mask Creation Software
The enhanced mask creation software has been used for the project. This enhanced software facilitates the creation of a much larger number of masks and has proved useful due to the acquisition of many more images for current and future projects. The refinements to the methodology allows for the masks to be created in a more streamlined fashion, and with better accuracy.
Pattern Classification Software
The pattern classification was modified for the additional images in project. Specifically, spectral features have been explored. Next is to apply the new texture features. Additionally, the leave-one-out testing technique has been integrated into CVIPtools.
Texture Features
Research has been undertaken for new texture features and the software is complete. It will be applied to the to the next Phase of the project, to the approximately 120 new COMS images. The new texture feature extraction software includes the five features that have been found to be the most useful: energy (homogeneity), inertia (contrast), correlation (linearity), inverse difference (local homogeneity) and entropy (details are in Spring 2009 report)
Algorithm Test and Analysis Tool
The development of the Algorithm Test and Analysis Tool (ATAT) should prove very useful as the number of projects expands, and the nature of the projects varies. Similar to what was done with the previous software development where we automated the experimental process so that we could run millions of experiments instead of hundreds, the ATAT software will allow for the investigation of many more permutations of image processing and feature extraction algorithms than was previously possible.
The ATAT software will allow the user to select from a set of image processing filters, segmentation methods, morphological filters, and various postprocessing imaging functions. The user will be able to specify the range and the increment to be used with each of the imaging functions’ parameters. The software will then automatically run all the possible algorithmic permutations. The success measure will be tailored to the specific application, and we are currently investigating the useful success metrics for the veterinary projects. An application manual for the current version of the software, CVIP-ATAT, is included in Appendix A of the Spring report.
Experimental Results and Discussion
Preliminary Experiments:
Preliminary experiments were performed previously to determine the optimal classification method and distance metric, optimal set of features, and best normalization method for the final experiments. Under consideration were the following:
CLASSIFICATION METHODS AND DISTANCE METRIC
1. Nearest Neighbor, distance metric: Euclidean
2. K-Nearest Neighbor with K = two through five, distance metric: Euclidean
FEATURES
1. Spectral features: sectors and rings, from 3 to 16.
2. Histogram features: mean,. Standard deviation, Skew, Energy and Entropy
3. Texture features: Energy, Inertia, Correlation, Inverse difference, and Entropy. The pixel distance was varied from two to sixteen.
DATA NORMALIZATION METHODS
1. Euclidean
2. Soft-max with r=1
Final Experiments:
After an initial set of experiments it was determined that following would be used for all the final experiments:
CLASSIFICATION METHOD AND DISTANCE METRIC
➢ K-Nearest Neighbor with K = 3, distance metric: Euclidean
FEATURES
➢ Histogram features: Mean. Standard deviation, Skew, Energy and Entropy
➢ Texture features: Energy, Inertia, Correlation, Inverse difference, and Entropy. The pixel distance was 6.
DATA NORMALIZATION METHOD
➢ Soft-max with r=1
The final experiments were performed with the 1,024 permutations of the ten features. 4096 experiments were performed.
Results from Experiment Set #1.
Images: Clinical group, the front and top images of the head.
Classes: moderate and severe.
Note: for complete results see the Excel spreadsheet files Experiment1_A1, Experiment1_A1D.
|Top five classification results for Experiment #1 |
|Features |Body Part |Number of images per class |Classification Success |
|(texture pixel dist=6) | | | |
|Texture energy |Top of head |Moderate: 79 |83% |
|Texture inverse diff | |Severe: 19 | |
|Histogram mean | | | |
|Histogram entropy | | | |
|Texture energy |Top of head |Moderate: 79 |82% |
| | |Severe: 19 | |
|Texture inverse diff |Front of head |Moderate: 79 |83% |
|Texture entropy | |Severe: 19 | |
|Histogram StdDev | | | |
|Histogram Skew | | | |
|Texture inverse diff |Front of head |Moderate: 79 |83% |
|Texture entropy | |Severe: 19 | |
|Histogram mean | | | |
|Histogram skew | | | |
|Texture inverse diff |Front of head |Moderate: 79 |83% |
|Texture entropy | |Severe: 19 | |
|Histogram StdDev | | | |
|Histogram skew | | | |
➢ We believe that the 83% classification success indicates that the top and front of head are the most useful for differentiation of the two classes.
➢ These results indicate that there is a difference between the moderate and severe classes in the clinical images.
➢ With the top of head image experiments 1 of the 1,024 feature sets returned an 83% classification rate.
➢ With the top of head image experiments the lowest classification rate was 67%.
➢ With the front of head image experiments 3 of the 1,024 feature sets returned a 83% classification rate
➢ With the front of head image experiments the lowest classification rate was 68%.
➢ The texture inverse difference and energy features appear the most predictive
Results from Experiment Set #1 – PCA Scatterplot
The principal components transform shows the clustering of the two classes. Here the original 25-dimensional feature space is linearly remapped to three dimensions in a manner that maximizes the retained information – in this case 79.2%.
This plot resulted from one of the 1,024 experiments in this set. The top of head image results are shown, and the classification success rate was 83%.
[pic]
Results from Experiment Set #2.
Images: Clinical group, the front and top images of the head.
Classes: mild, moderate and severe.
Note: for complete results see the Excel spreadsheet files Experiment2_A1, Experiment2_A1D, Experiment2_A1LL, Experiment2_A1LR.
|Top five classification results for Experiment #2 |
|Features |Body Part |Number of images per class |Classification Success |
|(texture pixel dist=6) | | | |
|Texture inertia |Top of head |Mild: 22 |65% |
|Texture correlation | |Moderate: 79 | |
|Texture inverse diff | |Severe: 19 | |
|Histogram mean | | | |
|Histogram StdDev | | | |
|Histogram energy | | | |
|Texture energy |Top of head |Mild: 22 |65% |
|Histogram mean | |Moderate: 79 | |
|Histogram skew | |Severe: 19 | |
|Histogram entropy | | | |
|Texture energy |Top of head |Mild: 22 |65% |
|Histogram mean | |Moderate: 79 | |
|Histogram skew | |Severe: 19 | |
|Histogram energy | | | |
|Histogram entropy | | | |
|Texture inertia |Front of head |Mild: 22 |65% |
|Texture correlation | |Moderate: 79 | |
|Texture inverse diff | |Severe: 19 | |
|Texture entropy | | | |
|Histogram std devHi | | | |
|Texture energy |Front of head |Mild: 22 |65% |
|Histogram skew | |Moderate: 79 | |
|Histogram entropy | |Severe: 19 | |
➢ These results are inconclusive but indicate that there may be a difference between the mild, moderate and severe classes in the clinical images.
➢ The highest classification rate achieved was 65%.
Results from Experiment Set #2 – PCA Scatterplot
The principal components transform shows the clustering of the three classes. Here the original xx-dimensional feature space is linearly remapped to three dimensions in a manner that maximizes the retained information – in this case 77.8%.
This plot resulted from one of the 1024 experiments in this set. The top of head images were used, and the classification success rate was 65%. As shown here, the mild class (shown in red) is niot easily separable from the moderate class (shown here in blue).
[pic]
Conclusions
Four sets of experiments were performed for a total for of 4096 experiments. The first set was to investigate classification of the moderate and severe classes with the clinical images. With top and front images, success rates of up to 83% were achieved. These results are not as good as the 100% previously achieved with smaller sets. This indicates that the class domains are more diverse, and the classes more correlated, than was found with the smaller set. Additionally, further experimentation with color normalization is warranted. The results do, however, still indicate that there is a difference between the moderate and severe classes with the clinical images. The features texture inverse difference and texture energy prove to be the most predictive features.
The second set of experiments used the clinical images and classified that were classified as severe, moderate and mild classes of the pathology. Here, the front and top images of the head were used. Success rates from 42% to 65% lead to inclusive results but indicate that there may be a difference between the mild, moderate and severe classes in the clinical images. We note a fair degree of overlap in the feature space between the mild and moderate classes. We believe this is what confuses the classifier, leading to a lower success rate.
Comparing the results from the first two sets of experiments we find that the two separate classes in the clinical (moderate and severe) images achieve better separation than with the three classes (mild, moderate and severe) in the clinical images.
Current and Future Projects Overview
Thermographic Evaluation of Elbow Dysplasia. Comparison of normal and abnormal thermographic images of the elbows was performed Summer 2009. Preliminary experiments have been performed and results are in that report.
Thermographic Evaluation of Feline Hyperthyroidism. Evaluation of normal feline thyroid and hyperthyroid glands thermographically. Assess the thermographic parameters of hyperthyroid cats thyroid glands pre and post radio-iodine treatment. Assess hyperthyroid cats thyroid glands pre and post clipping. Determine the optimal algorithm and imaging features for the classification.
Thermographic Imaging Characteristics of Type I Intervertebral Disc Disease in Dogs. To explore methods using infrared thermographic images and verify that they will correlate with findings of neurologic examinations, MRI and surgical findings in dogs with Type I intervertebral disc disease.
Thermographic imaging of cranial cruciate ligament tears of stifles in Labrador retrievers. To determine the accuracy of infrared thermographic imaging patterns in tears of cranial cruciate ligaments in Labrador retrievers. Infrared thermographic imaging findings will correlate with findings of pre-operative physical examination, and intra-operative findings in dogs with cranial cruciate ligament failure.
Thermographic evaluation of the canine stifle after cranial cruciate ligament repair and post-operative rehabilitation. Here we will compare the efficacy of three different post-operative rehabilitation methods: 1) electrical stimulation, 2) acupuncture, 3) cryotherapy. Thermographic images will be compared for the three different treatments by evaluation of the thermographs at various times and days post-op.
Further investigation for the Chiari malformation (COMS) project. More experiments with the new 120 canines can be perfomed with the color normalization methods (see Summer 2008 report), if the necessary information is supplied. This may improve the 83% success results achieved Fall 2009.
Future Tasks
• Continue improvement of the ATAT software
• Run more experiments for the Elbow Dysplasia project
• Further analyze results from the Elbow Dysplasia project
• Run the color normalization experiments for the COMS project with the new image set
• Run more experiments with the Cruciate Tears project, specifically the control, electro-stimulation, acupuncture and cryotherapy groups.
• Acquire images for the Disc Disease/Chronic Pain (Grossbard) project
• Design, run and analyze experiments for the Disc Disease/Chronic Pain (Grossbard) project
• Acquire images for the Feline Hyperthyroidism project
• Design, run and analyze experiments for the Feline Hyperthyroidism project
• Research the use of neural networks and genetic algorithms for the analysis of veterinary thermographic images
• Research methods to determine the “best” feature set(s)
• Enhance the graphical user interface (GUI) for the automatic pattern classification and application-specific software
• Explore finding of feature vector “norms” for various pathologies/abnormalities
• Explore multimodal image analysis
• Investigate use of support vector machines for analysis of veterinary thermographic images
References
1. Veterinary Thermographic Image Analysis, Project Number 7-64878, Report Number 4878-5, SIUE, August 20, 2008.
2. Veterinary Thermographic Image Analysis, Project Number 7-64878, Report Number 4878-6, SIUE, January 21, 2009.
3. Veterinary Thermographic Image Analysis, Project Number 7-64878, Report Number 4878-7, SIUE, May 13, 2009.
4. Veterinary Thermographic Image Analysis, Project Number 7-64878, Report Number 4878-8, SIUE, September 4, 2009.
5. Computer Imaging: Digital Image Analysis and Processing , Scott E Umbaugh, The CRC Press, Boca Raton, FL, January 2005
6. Pattern Recognition Engineering, Morton Adler and Eric Smith, John-Wiley & Sons, NY, 1993
7. Pattern Classification, 2nd Edition, Richard Duda, Peter Hart, David Stork, John-Wiley & Sons, NY, 2001
8. Pattern Recognition, Sergios Theodoris and Konstantinos Koutroumbas, Academic Press, NY, 2006
9. Catherine A. Loughin, Dominic J. Marino. “Evaluation of thermographic imaging of the limbs of healthy dogs.” American Journal of Veterinary Research, October 2007, Vol. 68, No. 10, Pages 1064-1069
10. Pead, Matthew J., and Sue Guthrie. "Elbow dysplasia in dogs - a new scheme." Elbow dysplasia in dogs - a new scheme (1997): 1-8. EE. SIUE. 12 July 2009
11. Prepared by the Orthopedic Foundation for Animals; “Examining Elbow Dysplasia”, Tufts Canine and Feline Breeding and Genetics Conference, 2005
12. Donna L. Harper, “The Value of Infrared Thermography in the Diagnosis and Prognosis of Injuries in Animals”, Proc., Inframation 2000, Orlando, USA (2000)
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related download
Related searches
- nevada board of veterinary examiners
- nevada board of veterinary medicine
- california board of veterinary examiners
- florida board of veterinary examiners
- hills veterinary diets
- state of nevada veterinary board
- pennsylvania veterinary license lookup
- georgia board of veterinary examiners
- ky board of veterinary examiners
- maryland board of veterinary examiners
- national board of veterinary examiners
- nevada veterinary license lookup