MIAD 2011 Abstracts


Full Papers
Paper Nr: 2
Title:

An Automated Tool for the Detection of Electrocardiographic Diagnostic Features based on Spatial Aggregation and Computational Geometry

Authors:

Liliana Ironi and Stefania Tentoni

Abstract: In this work we focus on Electrocardiographic diagnosis based on epicardial activation fields. The identification, within an activation map, of specific patterns that are known to characterize classes of pathologies provides an important support to the diagnosis of rhythm disturbances that can be missed by routine low resolution ECGs. Through an approach grounded on the integration of a Spatial Aggregation (SA) method with concepts borrowed from Computational Geometry, we propose a computational framework to automatically extract, from input epicardial activation data, a few basic features that characterize the wavefront propagation, as well as amore specific set of diagnostic features that identify an important class of rhythm pathologies due to block of conduction.

Paper Nr: 6
Title:

Features Extraction and Fuzzy Logic based Classification for False Positives Reduction in Mammographic Images

Authors:

Arianna Mencattini, Giulia Rabottino and Marcello Salmeri

Abstract: Breast cancer is one of the most common neoplasms in women and it is a leading cause of death worldwide. A proper screening procedure can help an early diagnosis of the tumor so reducing the death risk. A suitable computer aided detection system can help the radiologist to detect many subtle signs, normally missed during the screening phase, submitting to the radiologist’s attention several regions that could contain an abnormality. However, one of the most critical problem deals with a suitable tradeoff regarding the number of suspicious zones to present to the radiologist and the capability of identify the correct ones. In this work, the classification of suspicious signs into normal tissue or massive lesion has been faced in order to get a False Positive reduction without noticeably affecting the number of True Positives.

Paper Nr: 9
Title:

Breast Masses Classification using a Sparse Representation

Authors:

Fabián Narváez

Abstract: Breast mass detection and classification in mammograms is considered a very difficult task in medical image analysis. In this paper, we present a novel approach for classification of masses in digital mammograms according with their severity (benign or malign). Unlike other approaches, we do not segment masses but instead, we attempt to describe entire regions of interest (RoIs) based on a sparse representation. A set of patches selected by a radiologist in a RoI are characterized by their projection onto learned dictionaries, constructed previously from classified regions. Finally, the region class was identified using a decision rule algorithm. The strategy was assessed in a set of 80 masses with different shapes extracted from the DDSM database. The classification was compared with a ground truth already provided in the data base, showing an average accuracy rate of 70%.

Paper Nr: 12
Title:

Bone Surface Segmentation in Ultrasound Images: Application in Computer Assisted Intramedullary Nailing of the Tibia Shaft

Authors:

Agnès Masson-Sibut, Eric Petit, François Leitner and Julien Normand

Abstract: This paper deals with the use of ultrasound images in order to develop a Computer Assisted Orthopaedics Surgery system. Ultrasounds are easy to use in the Operating Room (OR), less expensive than other image modalities, and faster. We present an automatic method to extract anatomical landmarks from ultrasound images of femoral anterior condyles. The algorithm is based on an active contour model that uses an attraction field derived from an Euclidian-distance map. This segmentation process is a part of a global procedure that includes an interactive determination of the best image that could be chosen in order to obtain robust bone segmentation. This global procedure has been successfully tested on 11 volunteers.

Paper Nr: 16
Title:

3D Data Hiding for Enhancement and Indexation on Multimedia Medical Data

Authors:

N. Tournier and G. Subsol

Abstract: In medical applications, large quantities of multimedia data are exchanged such as 3D data acquired either by volume (CT-scan) or surface (laser range) scanning. The increasing of the numerical data using raises some unsolved issues. As for us, we are interested in the protection and the enhancement of multimedia content by insertion of hidden message. According to Koller et al. some of these challenges are: – Metadata embedding; – Indexing and searching in database. Data hiding may be a solution for these main applications in the medical domain. It is possible to embed metadata, with security for confidential data or for indexing area in a media, without increasing the size of the file.

Paper Nr: 17
Title:

Content-Based Computer Tomography Image Retrieval on a Whole-Body Anatomical Reference Set: Methods and Preliminary Results

Authors:

Auréline Quatrehomme and Denis Hoa

Abstract: This paper describes a CBIR system presenting two key points: a generic CT data, as well as a novel algorithm for combining visual features. The descriptors express grey levels, texture and shape of the images. A normalization method is proposed in order to improve the quality of indexing and retrieval. Our selected features and our combination method are effective for retrieving images from a whole-body reference set.

Paper Nr: 18
Title:

Current Challenges on Polyp Detection in Colonoscopy Videos: From Region Segmentation to Region Classification. A Pattern Recognition-based Approach

Authors:

Jorge Bernal

Abstract: In this paper we present our approach on selection of regions of interest in colonoscopy videos, which consists of three stages: Region Segmentation, Region Description and Region Classification, focusing on the Region Segmentation stage. As part of our segmentation scheme, we introduce our region merging algorithm that takes into account our model of appearance of the polyp. As the results show, the output of this stage reduces the number of final regions and indicates the degree of information of these regions. Our approach appears to outperform state-of-the-art methods. Our results can be used to identify polypcontaining regions in the later stages.

Short Papers
Paper Nr: 5
Title:

Image Segmentation Guidance using Pet Information on CT Images in PET/CT Dual Modality

Authors:

Iman Avazpour and Raja Syamsul Azmir Raja Abdullah

Abstract: Medical image segmentation has always relied on evaluation and processing of the target image. In this paper we are using PET/CT dual imaging modality data to start and guide segmentation of regions of interest on CT image. The aim is to improved current semi-automatic techniques to become fully automatics. The images are acquired for extra pulmonary tuberculosis (EPTB) to indicate the area of infections. Two segmentation algorithms have been examined and tested; Seeded Region Growing (SRG) and Watershed using this technique and their results have been evaluated considering segmentation accuracy and time consumption. Overall, adopting the proposed approach for boundary maximum gray value in SRG yields the best results in terms of the accuracy, and acceptable time computation.

Paper Nr: 10
Title:

Heritability Estimation Methods of Multiple Brain Measures: A Preliminary MRI Study in Twins

Authors:

Yu Yong Choi and Kun Ho Lee

Abstract: Heritability is the proportion of total phenotypic variance due to genetic influence. To estimate heritability, Falconer’s method and structural equation modeling (SEM) are used. However, compared to Falconer’s formula, SEM is hardly applicable for neuroimaging analysis because the SEM tools such as Mx cannot calculate numerous data simultaneously nor sequentially. We developed a code for multiple calculations using Mx to estimate the heritability of gray matter thickness at 81,924 surface points across the cerebral cortex. Although FF and SEM provided similar results, SEM was inclined to yield lower heritability estimates and more conservative significance than Falconer’s method. In considering the results, we propose that the correction for multiple comparisons should be carefully performed for the results from SEM.