Biomedical Image Classification with Random Subwindows and Decision Trees

Abstract

In this paper, we address a problem of biomedical image classification that involves the automatic classification of x-ray images in 57 predefined classes with large intra-class variability. To achieve that goal, we apply and slightly adapt a recent generic method for image classification based on ensemble of decision trees and random subwindows [MGPW05]. We obtain classification results close to the state of the art on a publicly available database of 10000 x-ray images. We also provide some clues to interpret the classification of each image in terms of subwindow relevance.

Random subwindows and extremely randomized trees for image classification in cell biology

“…Moreover, in [21], we successfully applied it on a 10000 X-Ray image database with classification results very close to the best ones [13]. …”

Medical Image Annotation and Retrieval Using Visual Features

“…Since the scale and rotation variations in radiographs of the same category are small, the SIFT descriptor can not show its advantage of being scale and rotation invariant for describing radiographs. In previous works, local image patches have shown pleasant performance for medical image retrieval and classification [5][6] [7]. Therefore, we utilize local image patches as the local features in our experiments.…”

Using texture-based symbolic features for medical image representation

“…The IRMA project proposes a general structure for semantic medical image analysis [9], and recently, bodyregion categorization results are presented, taking in consideration multiple modalities, but focusing on X-Rays [7]. On the same data set, [11] present another classification approach, based on the extraction of random sub-windows from X-Ray images, and their classification with decision trees. Recently, a modality and anatomical region categorization evaluation was performed using the MedIC module and the MedGIFT 1 system [3].…”