Automatic small bowel tumor diagnosis by using multi-scale wavelet-based analysis in wireless capsule endoscopy images

Abstract

BackgroundWireless capsule endoscopy has been introduced as an innovative, non-invasive diagnostic technique for evaluation of the gastrointestinal tract, reaching places where conventional endoscopy is unable to. However, the output of this technique is an 8 hours video, whose analysis by the expert physician is very time consuming. Thus, a computer assisted diagnosis tool to help the physicians to evaluate CE exams faster and more accurately is an important technical challenge and an excellent economical opportunity.MethodThe set of features proposed in this paper to code textural information is based on statistical modeling of second order textural measures extracted from co-occurrence matrices. To cope with both joint and marginal non-Gaussianity of second order textural measures, higher order moments are used. These statistical moments are taken from the two-dimensional color-scale feature space, where two different scales are considered. Second and higher order moments of textural measures are computed from the co-occurrence matrices computed from images synthesized by the inverse wavelet transform of the wavelet transform containing only the selected scales for the three color channels. The dimensionality of the data is reduced by using Principal Component Analysis.ResultsThe proposed textural features are then used as the input of a classifier based on artificial neural networks. Classification performances of 93.1% specificity and 93.9% sensitivity are achieved on real data. These promising results open the path towards a deeper study regarding the applicability of this algorithm in computer aided diagnosis systems to assist physicians in their clinical practice.

Gait analysis parameters of healthy human subjects with asymmetric loads

“…Tumor diagnosis based on processing endoscopic images with wavelets is proposed by Barbosa et al (2012). A review of time-frequency analysis methods (Fourier Transform, DWT) usable for circadian and ultradian patterns in behavioral records is made by Leise (2013).…”

Evaluation and management of small-bowel tumors in the era of deep enteroscopy

“…45 Other attempts to differentiate tumors on CE include an automated scale that uses wavelet-based analysis in CE images; this has a reported 95% sensitivity and specificity. 46 Additionally, efforts have been made to identify major criteria for a high-probability mass lesion on CE (major criteria: bleeding, mucosal disruption, irregular surfaces, polypoid appearance, color). 47 CE has several other limitations in the evaluation of small-bowel tumors.…”

The Current Main Types of Capsule Endoscopy

“…Small bowel tumors can be easily missed due to the predominant submucosal and extraluminal location of the tumors. Specific indexes and scales were developed for improving the detection rate of small bowel tumors, including the Smooth Protruding Index on Capsule Endoscopy (SPICE score) and an automated scale using multiscale wavelet-based analysis [23,24].…”