M.E. Jernigan SB, SM, PhD (MIT), PEng
Distinguished Teacher's Award, 1986
Office: DC 3601
Phone: (519) 888-4567 ext. 84644
Fax: (519) 746-4791
Email: ![]()
Recently, researchers in vision and image processing have developed quantitative models of both image formation and human perceptual performance which provide new insight into the form and function of image processing and pattern recognition tasks in vision. Research in image processing in Systems Design Engineering is part of an emerging effort to develop a coherent framework for image and signal processing and analysis based on perceptual models. Two current specific objectives are the development of image enhancement, segmentation and classification systems using non-linear and adaptive filters suited to non-additive, non-stationary noise and signal processes characteristic of imaging systems, and development of machine perception algorithms for texture and shape. Applications include enhancement and restoration for remote sensing (SAR) and medical (ultrasound, MRI, PET) images.
This research is concerned with providing computers with a perceptual capability in the sense that the significant underlying patterns in a given data set are to be automatically classified or recognized. Data analysis is approached as a problem in determining the similarity between an unknown input pattern and known classes represented in memory.
Both single and multidimensional signal processing problems are considered within the pattern recognition framework. One dimensional applications include speech processing and eye movement analysis in visual diagnosis. Visual patterns presented as images are described with texture measures for image segmentation, classification and interpretation. Ongoing projects include texture segmentation of synthetic aperture radar (SAR) imagery of sea ice in remote sensing applications and medical image analysis.