10 - Improved extraction of the extreme thermal regions of breast IR images

Abstract

Early detection of breast tumors has many clinical advantages such as treating primary tumors without surgery, considering more treatment options. Psychologically it is more approachable to the patients since their chance of survival will be increased. Thermography is a technique having potential to detect breast abnormalities earlier than other methods. Hot areas in breast thermal images have susceptibility to be suspicious. Accordingly, segmenting the extreme thermal areas in breast thermal images is valuable. Lazy snapping is an interactive image cutout procedure that can be used to segment the extreme thermal spots in breast thermograms rapidly with effortless detailed transformation. The most essential benefit of this procedure is that it can afford the results for medical professionals in real time promptly. By way of explanation, it is a valuable interactive image segmentation method because of its basic characteristics. These are as follows: firstly, the technique concludes intuitive segmentation that replies the user attempt with given a definite user input and secondly, the technique has sufficient power to afford quick evaluation. In spite of different procedures such as fuzzy c means, level set, K means, and mean shift procedures were used by the authors to extract extreme thermal regions in breast thermograms in their previous studies, lazy snapping procedure was more user-friendly and can allow quick assessment. In this research, nine thermal cases were demonstrated and by implementing lazy snapping procedure, the extreme thermal areas were segmented from the provided breast data set. For the nine presented cases, the time taken to extract the extreme thermal regions varied from less than 5 to 10s. It was found that lazy snapping procedure was quicker than other techniques used by the authors for extracting extreme thermal regions in female breast thermal images.