Automated Growcut for segmentation of endoscopic images

Multiple abnormality detection for automatic medical image diagnosis using bifurcated convolutional neural network

“…Also, utilizing primary parts of the network prevented redundant operations which makes it suitable for implementation inside the portable medical devices. --Sen: 0.88 Spec : 0.84 ------SVM [26] --Sen : 0.78 Spec : 0.99 ------SVM [9] ----DICE: 0.84 ----SVM [8] ----DICE: 0.81 ----ANN [40] ----DICE: 0.85 ----SVM [10] ----AUC: 0.83 AUC : 0.87 AUC: 0.96 CNN [27] ----…”

“…Finally, superpixels are classified using a support vector machine (SVM) approach. In [8], initial seeds detected using K-means are labeled by SVM, then bleeding regions are segmented using an interactive segmentation based on cellular automata. In [9], WCE images are represented using local binary pattern and average saturation and are classified using SVM in bleeding and non-bleeding regions.…”

“…In [7], [25] and [26], SVM is utilized for segmentation of bleeding and angioectasias images. In [8] and [9], bleeding WCE images are segmented using SVM. Also in [40], ANN is used for segmentation of bleeding images.…”