Application of 3D-Representation Algorithms in Small-Bowel Capsule Endoscopy

Abstract

In video endoscopy, there is on-going research for a three-dimensional (3-D) reconstruction of the gastrointestinal (GI) tract using stereoscopic cameras [1,2]. This could lead to improved diagnostic yield by enhancing textural features of the mucosa and/or abnormalities [3,4]. However, due to current limitations in capsule endosocpy (CE) technology i.e. size, power consumption, and resolution, accurate 3-D imaging of the intestinal lumen in CE is still unfeasible [5]. To address this shortcoming, various algorithms such as Shape-from-Shading (SfS) have been utilized in CE. Essentially, SfS algorithms reconstruct surface appearances through an inversion of the image irradiance equation thus recovering the shape of GI structures and/or lesions given a single twodimensional 2-D image. During a recent CE examination in our Department, the ampulla of Vater was identified (Figure 1). By applying 4 publicly available SfS algorithms [3, 6-9] on the 2-D image surface features become accentuated (Figure 2), although the outcome is variable and depends on the method used (unpublished authors' data).