Factors influencing the accuracy for tissue classification in multi spectral in-vivo endoscopy for the upper gastro-internal tract.

Hyper spectral imaging is a possible way for disease detection. However, for carcinoma detection most of the results are ex-vivo. However, in-vivo results of endoscopic studies still show fairly low accuracies in contrast to the good results of many ex-vivo studies. To overcome this problem and to provide a reasonable explanation, Monte-Carlo simulations of photon trajectories are proposed as a tool to generate multi spectral images including inter patient variations to simulate 40 patients. Furthermore, these simulations have the huge advantage that the position of the carcinoma is known. Due to this, the effect of mislabelled data can be studied. As shown in this study, a percentage of 30-35% of mislabelled data might lead to significant decrease of the accuracy from around 90% to around 70-75%. Therefore, the main focus of hyper spectral imaging has to be the exact characterization of the training data in the future.

Wireless capsule endoscopy color video segmentation.

This paper describes the use of color image analysis to automatically discriminate between oesophagus, stomach, small intestine, and colon tissue in wireless capsule endoscopy (WCE). WCE uses "pill-cam" technology to recover color video imagery from the entire gastrointestinal tract. Accurately reviewing and reporting this data is a vital part of the examination, but it is tedious and time consuming. Automatic image analysis tools play an important role in supporting the clinician and speeding up this process. Our approach first divides the WCE image into subimages and rejects all subimages in which tissue is not clearly visible. We then create a feature vector combining color, texture, and motion information of the entire image and valid subimages. Color features are derived from hue saturation histograms, compressed using a hybrid transform, incorporating the discrete cosine transform and principal component analysis. A second feature combining color and texture information is derived using local binary patterns. The video is segmented into meaningful parts using support vector or multivariate Gaussian classifiers built within the framework of a hidden Markov model. We present experimental results that demonstrate the effectiveness of this method.