The Classification of Endoscopy Images with Persistent Homology

Aiming at the automatic diagnosis of tumors from narrow band imaging (NBI) magnifying endoscopy (ME) images of the stomach, we combine methods from image processing, computational topology, and machine learning to classify patterns into normal, tubular, vessel. Training the algorithm on a small number of images of each type, we achieve a high rate of correct classifications. The analysis of the learning algorithm reveals that a handful of geometric and topological features are responsible for the overwhelming majority of decisions.

[1]  Milan Sonka,et al.  Image pre-processing , 1993 .

[2]  K. Yao,et al.  A New Diagnostic VS Classification System Produced by Magnification Endoscopy Plus Narrow-Band Imaging in the Stomach: Microvascular Architecture and Microsurface Structure , 2008 .

[3]  Robert E. Tarjan,et al.  Data structures and network algorithms , 1983, CBMS-NSF regional conference series in applied mathematics.

[4]  Nuno Vasconcelos,et al.  Endoscopic image analysis in semantic space , 2012, Medical Image Anal..

[5]  James R. Munkres,et al.  Elements of algebraic topology , 1984 .

[6]  Shoichi Saito,et al.  Imaging by Magnifying Endoscopy with NBI Implicates the Remnant Capillary Network As an Indication for Endoscopic Resection in Early Colon Cancer , 2011, International journal of surgical oncology.

[7]  Til Aach,et al.  Classification of colon polyps in NBI endoscopy using vascularization features , 2009, Medical Imaging.

[8]  Sariel Har-Peled.  Geometric Approximation Algorithms , 2011 .

[9]  Afra Zomorodian,et al.  Computational topology , 2010 .

[10]  Tohru Ishitani,et al.  Contrast-to-gradient method for the evaluation of image resolution in scanning electron microscopy , 2002 .

[11]  Andreas Uhl,et al.  Predicting the histology of colorectal lesions in a probabilistic framework , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Workshops.

[12]  K. Chayama,et al.  Quantitative analysis and development of a computer-aided system for identification of regular pit patterns of colorectal lesions. , 2010, Gastrointestinal endoscopy.

[13]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[14]  Andreas Uhl,et al.  Endoscopic Image Classification Using Edge-Based Features , 2010, 2010 20th International Conference on Pattern Recognition.

[15]  F. Wrba,et al.  Pit pattern classification using extended Local Binary Patterns , 2009, 2009 9th International Conference on Information Technology and Applications in Biomedicine.

[16]  Herbert Edelsbrunner,et al.  The Adaptive Topology of a Digital Image , 2012, 2012 Ninth International Symposium on Voronoi Diagrams in Science and Engineering.

[17]  William M. Tierney,et al.  Narrow band imaging and multiband imaging. , 2008, Gastrointestinal endoscopy.

[18]  Daniel P. Huttenlocher,et al.  Distance Transforms of Sampled Functions , 2012, Theory Comput..