Superpixel-based automatic segmentation of villi in confocal endomicroscopy

Abstract

Confocal Laser Endomicroscopy (CLE) is a technique permitting on-site microscopy of the gastrointestinal mucosa after the application of a fluorescent agent, allowing the evaluation of mucosa alterations. These are used as features by skilled technicians to stage the severity of multiple diseases, celiac disease or irritable bowel syndrome among the others. We present an automatic method for villi detection from confocal endoscopy images, whose appearance changes with mucosal alterations. Superpixel segmentation, a well-known technique originating from computer vision, is used to identify and cluster together pixels belonging to uniform regions. Each image in the dataset is analyzed in a multiscale fashion (scale 1, 0.5 and 0.25). From each superpixel, 37 features are extracted at multiple image scales. Each superpixel is classified using a random forest, and a post-processing step is performed to refine the final output. Results in the test set (70 images, 30870 superpixels) show 85.87% accuracy, 92.88% sensitivity, 76.99% specificity in the superpixel space, and 86.36% of accuracy and 87.44% Dice score in the pixel domain.

Impact of image segmentation techniques on celiac disease classification using scale invariant texture descriptors for standard flexible endoscopic systems

“…At this phase, segmentation techniques can be used along with image noise reduction and enhancement methods, and this was done in our study. Boschetto et al, [10]. As a result of the increasing demand in computer visual systems in recent years, image thresholding has been continuing, which has attracted the attention of researchers in different fields.…”

Image computing for fibre-bundle endomicroscopy: A review

“…(Boschetto et al, 2015a;Boschetto et al, 2016b;Boschetto et al, 2015b) attempted to semi-automatically analyse and quantify fluorescent endomicroscopic images of the gastro-intestinal mucosa, as a first step to assist diagnosis and monitoring of Celiac Disease. (Boschetto et al, 2016b;Boschetto et al, 2015b) proposed methodologies for segmenting intestinal villi, while (Boschetto et al, 2015a) proceeded in detecting and segmenting cells within the villi and differentiating between columnar and goblet cells of the epithelium.…”

“…In the orapharyngeal tract,(Mualla et al, 2014) identified the borders and locations respectively of epithelial cells in the mucosa layer of vocal chords as the first step to analysing and quantifying structural changes.In the gastrointestinal tract,(Couceiro et al, 2012) developed a methodology that employed off-the-shelf algorithms for segmenting and quantifying intestinal crypts in endomicroscopic images as a potential indicator for Inflammatory Bowel Disease. Similarly,(Prieto et al, 2016) employed crypt detection as a first step towards automated classification between benign and dysplastic epithelial tissue in colorectal polyps (Boschetto et al, 2015a;Boschetto et al, 2016b;Boschetto et al, 2015b). attempted to semi-automatically analyse and quantify fluorescent endomicroscopic images of the gastro-intestinal mucosa, as a first step to assist diagnosis and monitoring of Celiac Disease (Boschetto et al, 2016b;Boschetto et al, 2015b).…”

“…Similarly,(Prieto et al, 2016) employed crypt detection as a first step towards automated classification between benign and dysplastic epithelial tissue in colorectal polyps (Boschetto et al, 2015a;Boschetto et al, 2016b;Boschetto et al, 2015b). attempted to semi-automatically analyse and quantify fluorescent endomicroscopic images of the gastro-intestinal mucosa, as a first step to assist diagnosis and monitoring of Celiac Disease (Boschetto et al, 2016b;Boschetto et al, 2015b). proposed methodologies for segmenting intestinal villi, while(Boschetto et al, 2015a) proceeded in detecting and segmenting cells within the villi and differentiating between columnar and goblet cells of the epithelium.Finally, in the pulmonary tract,(Namati et al, 2008) analysed mice distal lung images and automatically quantified the number and size of alveolar sacs.…”