Albert Torrent

Automatic microcalcification and cluster detection for digital and digitised mammograms

In this paper we present a knowledge-based approach for the automatic detection of microcalcifications and clusters in mammographic images. Our proposal is based on using local features extracted from a bank of filters to obtain a local description of the microcalcifications morphology. The developed approach performs an initial training step in order to automatically learn and select the most salient features, which are subsequently used in a boosted classifier to perform the detection of individual microcalcifications. Subsequently, the microcalcification detection method is extended in order to detect clusters. The validity of our approach is extensively demonstrated using two digitised databases and one full-field digital database. The experimental evaluation is performed in terms of ROC analysis for the microcalcification detection and FROC analysis for the cluster detection, resulting in better than 80% sensitivity at 1 false positive cluster per image.

Breast Density Segmentation: A Comparison of Clustering and Region Based Techniques

This paper presents a comparison of two clustering based algorithms and one region based algorithm for segmenting fatty and dense tissue in mammographic images. This is a crucial step in order to obtain a quantitative measure of the density of the breast. The first algorithm is a multiple thresholding algorithm based on the excess entropy, the second one is based on the Fuzzy C-Means clustering algorithm, and the third one is based on a statistical analysis of the breast. The performance of the algorithms is exhaustively evaluated using a database of full-field digital mammograms containing 150 CC and 150 MLO images and ROC analysis (ground-truth provided by an expert). Results demonstrate that the use of region information is useful to obtain homogeneous region segmentation, although clustering algorithms obtained better sensitivity.

Automatic Diagnosis of Masses by Using Level set Segmentation and Shape Description

We present here an approach for automatic mass diagnosis in mammographic images. Our strategy contains three main steps. Firstly, region of interests containing mass and background are segmented using a level set algorithm based on region information. Secondly, the characterisation of each segmented mass is obtained using the Zernike moments for modelling its shape. The final step is the diagnosis of masses as benign or malignant lesions, which is done using the Gentleboost algorithm that also assigns a likelihood value to the final result. The experimental evaluation, performed using two different digitised databases and Receiver Operating Characteristics (ROC) analysis, proves the feasibility of our proposal, showing the benefits of a correct shape description for improving automatic mass diagnosis.

MammoApplet: An interactive Java applet tool for manual annotation in medical imaging

Web-based applications in computational medicine have become increasingly important during the last years. The rapid growth of the World Wide Web supposes a new paradigm in the telemedicine and eHealth areas in order to assist and enhance the prevention, diagnosis and treatment of patients. Furthermore, training of radiologists and management of medical databases are also becoming increasingly important issues in the field. In this paper, we present MammoApplet, an interactive Java applet interface designed as a web-based tool. It aims to facilitate the diagnosis of new mammographic cases by providing a set of image processing tools that allow a better visualization of the images, and a set of drawing tools, used to annotate the suspicious regions. Each annotation allows including the attributes considered by the experts when issuing the final diagnosis. The overall set of overlays is stored in a database as XML files associated with the original images. The final goal is to obtain a database of already diagnosed cases for training and enhancing the performance of novice radiologists.

Segmenting extended structures in radio astronomical images by filtering bright compact sources and using wavelets decomposition

The automatic segmentation of extended real structures (such as SNRs, HII regions, bow shocks, etc) in large surveys is a difficult task due to their morphological complexity and their wide variety in scale and surface brightness. With the aim of dealing with these issues we propose in this paper an automatic segmentation method based on applying wavelet decomposition in the residual thresholded images. The use of this strategy, instead of a wavelet decomposition on the original image, allows to avoid the artifacts produced by strong sources. Experimental results with radio images demonstrate that the proposed method successfully segments extended structures at different scales and therefore is suitable for further morphological analysis and object recognition processes.

A boosting based approach for automatic micro-calcification detection

In this paper we present a boosting based approach for automatic detection of micro-calcifications in mammographic images Our proposal is based on using local features extracted from a bank of filters for obtaining a description of the different micro-calcifications morphology The approach performs an initial training step in order to automatically learn and select the most salient features, which are subsequently used in a boosting classifier to perform the detection The validity of our method is demonstrated using 112 mammograms of the well-known digitised MIAS database and 280 mammograms of a full-field digital database The experimental evaluation is performed in terms of ROC analysis, obtaining Az=0.88 and Az=0.90 respectively, and FROC analysis The obtained results show the feasibility of our approach for detecting micro-calcifications in both digitised and digital technologies.

One-shot segmentation of breast, pectoral muscle, and background in digitised mammograms

The segmentation of the breast from the background and the pectoral muscle is the first pre-processing step in computerised mammographic analysis. This problem is usually solved by dividing it into two different segmentation strategies, one for the background and another one for the pectoral muscle. In this paper we tackle this problem jointly using a supervised single strategy. Namely, from a set of manually segmented mammograms, we model each of the three regions (breast, pectoral muscle, and background) using position, intensity, and texture information. Although the approach requires a training step, it allows a fast and reliable segmentation of new mammograms. The obtained results using 149 mammograms of the MIAS database show a high degree of overlap between manual and automatic segmentation.

Simultaneous detection and segmentation for generic objects

Numerous approaches to object detection and segmentation have been proposed so far. However, these methods are prone to fail in some general situations due to the proper object nature. For instance, classical approaches of object detection and segmentation obtain good results for some specific object classes (i.e. detection of pedestrians or segmentation of cars). However, these methods have troubles when detecting or segmenting object classes with different distinctive characteristics (i.e. cars and horses versus sky and road). In this paper, we propose a general framework to simultaneously perform object detection and segmentation on objects of different nature. Our approach is based on a boosting procedure which automatically decides - according to the object properties - whether is better to give more weight to the detection or segmentation process to improve both results. We validate our approach using different object classes from La-belMe, TUD and Weizmann databases, obtaining competitive detection and segmentation results.

Detecting Faint Compact Sources Using Local Features and a Boosting Approach

Several techniques have been proposed so far in order to perform faint compact source detection in wide field interferometric radio images. However, all these methods can easily miss some detections or obtain a high number of false positive detections due to the low intensity of the sources, the noise ratio, and the interferometric patterns present in the images. In this paper we present a novel strategy to tackle this problem. Our approach is based on using local features extracted from a bank of filters in order to provide a description of different types of faint source structures. We then perform a training step in order to automatically learn and select the most salient features, which are used in a Boosting classifier to perform the detection. The validity of our method is demonstrated using 19 real images that compose a radio mosaic. The comparison with two well-known state of the art methods shows that our approach is able to obtain more source detections, reducing also the number of false positives.

A supervised micro-calcification detection approach in digitised mammograms

We present in this paper a supervised approach for automatic detection of micro-calcifications. The system is based on learning the different morphology of the micro-calcifications using local features, which are extracted using a bank of filters. Afterwards, this set of features is used to train a pixel-based boosting classifier which at each round automatically selects the most salient one. Therefore, when a new mammogram is tested only the salient features are computed and used to classify each pixel of the mammogram as being part of a micro-calcification or actually being normal tissue. The experimental results shows the validity of our approach. Moreover, the robustness of our method is also demonstrated using a digitised database for the learning process and a different one for the testing, providing satisfactory results.