[Retracted] Using Convolutional Neural Networks for Segmentation of Multiple Sclerosis Lesions in 3D Magnetic Resonance Imaging

Figure 6

Example of an image in which the automatic segmentation of WM fails. (a) Preprocessed T1 image. (b) Preprocessed T2 image. (c) Preprocessed FLAIR image. (d) WM obtained from K-means. (e) Segmentation delivered by the CHB expert. (f) Segmentation obtained from the GMM. (g) Segmentation obtained by CNN (without max pooling). (h) Segmentation obtained by CNN (with Max pooling). (i) CNN multiview segmentation (no Max pooling). (j) CNN multiview segmentation (with Max pooling).