Title: Efficient photometric feature extraction in a hierarchical learning scheme for nodule detection
Authors: Ömer Muhammet Soysal; Jianhua Chen; Helmut Schneider
Addresses: Department of Information Systems and Decision Sciences, Louisiana State University, Baton Rouge, LA 70803-4020, USA. ' Department of Computer Science, Louisiana State University, Baton Rouge, LA 70803-4020, USA. ' Department of Information Systems and Decision Sciences, Louisiana State University, Baton Rouge, LA 70803-4020, USA
Abstract: In this paper, we present an efficient way of computing a run-length matrix which is used in extracting photometric features for computer vision applications such as object recognition and image retrieval. Our algorithm has two main steps which are quantisation and a fast indexing procedure for grey levels and their runs. We tested the features computed from the run-length matrix in our computer aided detection (CAD) system for lung nodule detection from computed tomography (CT) images. Our CAD system embeds a hierarchical learning scheme that allows multi-perspective and multi-level object recognition. The classification results obtained using the run-length features are promising.
Keywords: run-length features; RLF; computer aided detection; object recognition; image retrieval; texture features; computed tomography; content-based image representation; granular computing; intelligent systems; modular learning; nodule detection; rough sets; photometric features; computer vision; feature extraction; hierarchical learning; lung nodules; CT images; lung cancer; medical imaging.
International Journal of Granular Computing, Rough Sets and Intelligent Systems, 2012 Vol.2 No.4, pp.314 - 326
Available online: 21 Oct 2012 *Full-text access for editors Access for subscribers Purchase this article Comment on this article