Scott E. Umbaugh

An automatic color segmentation algorithm with application to identification of skin tumor borders

A principal components transform algorithm for automatic color segmentation of images is described. This color segmentation algorithm was used to find tumor borders in six different color spaces including the original red, green, and blue (RGB) color space of the digitized image, the intensity/hue/saturation (IHS) transform, the spherical transform, chromaticity coordinates, the CIE transform and the uniform color transform designated CIE-LUV. Five hundred skin tumor images were separated into a training set and a test set for comparison of the different color spaces. Automatic induction was applied to dynamically determine the number of colors for segmentation. Ninety-one percent of image variance was contained in the image component along the principal axis (also containing the most image information). When compared to a luminance radial search method, the principal components color segmentation border method performed equally well by one measure and 10% better by another measure, including more near border points outside the tumor. The spherical transform provides the highest success rate and the chromaticity transform the lowest error rate, although large variances in the data preclude definitive statistical comparisons.

Feature extraction in image analysis. A program for facilitating data reduction in medical image classification

Images are important for many biomedical applications. Here, the authors focus on the feature-extraction part of the image analysis process. The following topics are dealt with: feature vectors and feature spaces; binary object features; histogram features; color features; spectral features; feature extraction using CVIPtools; analysis/preprocessing.

Skin lesion classification using relative color features

Background/purpose: Clinically, it is difficult to differentiate the early stage of malignant melanoma and certain benign skin lesions due to similarity in appearance. Our research uses image analysis of clinical skin images and relative color‐based pattern recognition techniques to enhance the images and improve differentiation of these lesions.

Detection of atypical texture features in early malignant melanoma

Background: The presence of an atypical (irregular) pigment network (APN) can indicate a diagnosis of melanoma. This study sought to analyze the APN with texture measures.

Performance of AI methods in detecting melanoma

This research has shown that features extracted from color skin tumor images by computer vision methods can be reliable discriminators of malignant tumors from benign ones. Reliability was demonstrated by the monotonically increasing success ratios with increasing training set size and by the small standard deviations from the mean success rates. An average success rate of 70 percent in diagnosing melanoma was attained for a training set size of 60 percent. The presence or absence of atypical moles in the training and test sets was shown to have a dramatic impact on the effectiveness of the generated classification rules. This was the case with both AIM and lst-Class, and indicates a high potential for success if a method can be found for discriminating between atypical moles and melanoma. >

Nondermatoscopic digital imaging of pigmented lesions

Background/aims: Pigmented lesions are often difficult to evaluate clinically. Improvement of diagnostic accuracy by dermatoscopy has attracted much interet. With advanced digital imaging measurement of assymmetry, border irregularity and relative color as well as texture characteristics, lesional depth and changes in lesional area are now possible, the object of this review is to conclude the present status of these techniques and their potential.

Thermal imaging of normal and cranial cruciate ligament-deficient stifles in dogs.

Objective: To investigate the capability of thermography for differentiation between normal stifles and those with cranial cruciate ligament (CCL) rupture in dogs, initially with a full hair coat and 1 hour after clipping the hair coat. Study Design: Prospective study. Animals: Labrador Retrievers (n=6) with normal stifle joints (controls) and adult dogs (n=10) with CCL rupture. Methods: Thermography was performed before, and 60 minutes after, clipping the hair coat from the pelvic limb. Stifle images were classified as normal or abnormal, then subclassified as clipped and unclipped hair coat. CCL deficiency was confirmed at surgery and thermographic images subsequently classified as abnormal before analysis with image processing software. Results: Using image recognition analysis, differentiation between normal and CCL-deficient stifles in both clipped and unclipped dogs was 85% successful on cranial images, medial, caudal, and lateral images were between 75% and 85% successful. Although there were significant increases in skin temperature after clipping in both groups (P<.0002–.0001), there were no significant temperature differences between normal and CCL-deficient stifles when the entire stifle was examined. Conclusion: Thermography was successful in differentiating naturally occurring CCL-deficient stifles in dogs, with a success rate of 75–85%. Clipping is not necessary for successful thermographic evaluation of the canine stifle. Clinical Relevance: Thermography may be a useful imaging modality for diagnosis of CCL deficiency in dogs when CCL rupture is suspected but stifle laxity is not evident.OBJECTIVE To investigate the capability of thermography for differentiation between normal stifles and those with cranial cruciate ligament (CCL) rupture in dogs, initially with a full hair coat and 1 hour after clipping the hair coat. STUDY DESIGN Prospective study. ANIMALS Labrador Retrievers (n=6) with normal stifle joints (controls) and adult dogs (n=10) with CCL rupture. METHODS Thermography was performed before, and 60 minutes after, clipping the hair coat from the pelvic limb. Stifle images were classified as normal or abnormal, then subclassified as clipped and unclipped hair coat. CCL deficiency was confirmed at surgery and thermographic images subsequently classified as abnormal before analysis with image processing software. RESULTS Using image recognition analysis, differentiation between normal and CCL-deficient stifles in both clipped and unclipped dogs was 85% successful on cranial images, medial, caudal, and lateral images were between 75% and 85% successful. Although there were significant increases in skin temperature after clipping in both groups (P<.0002-.0001), there were no significant temperature differences between normal and CCL-deficient stifles when the entire stifle was examined. CONCLUSION Thermography was successful in differentiating naturally occurring CCL-deficient stifles in dogs, with a success rate of 75-85%. Clipping is not necessary for successful thermographic evaluation of the canine stifle. CLINICAL RELEVANCE Thermography may be a useful imaging modality for diagnosis of CCL deficiency in dogs when CCL rupture is suspected but stifle laxity is not evident.

Digital Image Processing and Analysis: Human and Computer Vision Applications with CVIPtools, Second Edition

Whether for computer evaluation of otherworldly terrain or the latest high definition 3D blockbuster, digital image processing involves the acquisition, analysis, and processing of visual information by computer and requires a unique skill set that has yet to be defined a single text. Until now. Taking an applications-oriented, engineering approach, Digital Image Processing and Analysis provides the tools for developing and advancing computer and human vision applications and brings image processing and analysis together into a unified framework.Providing information and background in a logical, as-needed fashion, the author presents topics as they become necessary for understanding the practical imaging model under study. He offers a conceptual presentation of the material for a solid understanding of complex topics and discusses the theory and foundations of digital image processing and the algorithm development needed to advance the field. With liberal use of color through-out and more materials on the processing of color images than the previous edition, this book provides supplementary exercises, a new chapter on applications, and two major new tools that allow for batch processing, the analysis of imaging algorithms, and the overall research and development of imaging applications. It includes two new software tools, the Computer Vision and Image Processing Algorithm Test and Analysis Tool (CVIP-ATAT) and the CVIP Feature Extraction and Pattern Classification Tool (CVIP-FEPC).Divided into five major sections, this book provides the concepts and models required to analyze digital images and develop computer vision and human consumption applications as well as all the necessary information to use the CVIPtools environment for algorithm development, making it an ideal reference tool for this fast growing field.

Compression of skin tumor images

The following topics are dealt with: error measures; transform-domain processing; preprocessing of color images; the principal components transform; the discrete wavelet transform; wavelet decomposition and reconstruction; vector quantization; distortion measures; a codebook design algorithm; design and implementation of compression algorithms; preprocessing effects; quality assessment of the compressed images.

Differentiation of melanoma from benign mimics using the relative-color method.

Background: Previous studies have successfully classified 86% of malignant melanomas using a relative‐color segmentation method, by feature extraction from photographic images in the automatic identification of skin tumors. These studies were extended by applying the relative‐color method to dermoscopic images of melanoma grouped with melanoma in situ and clark nevus lesions in dermoscopic images allow more control over lighting variations, which contribute to lesion misclassification. Dermoscopic images then enable a more detailed examination of the structure of skin lesions, provide much more structural detail within lesions, and contain visual information that cannot be seen in photographic images. This present work extends the previous studies by applying relative‐color feature extraction to dermoscopic images to differentiate among melanoma, seborrheic keratoses and Reed/Spitz nevi.