9 research outputs found
Characterization of an FFDM unit based on a-Se direct conversion detector
No full textThe aim of this paper is to investigate the properties of a clinical FFDM unit (Giotto-Image MD, IMS Italy). The digital detector consists of a flat panel using the amorphous selenium technology (ANRAD Corporation, Canada). The active area of the imager is 17.4 cm x 23.9 cm (2048 x 2816 pixels) with a pixel pitch of 85 gm. The direct conversion of X-rays into charge provides excellent imaging performance. In this work we present an objective and complete characterization of such system: detector response, MTF, NPS and DQE calculation will be presented. MTF and DQE at Nyquist frequency (5.88 lp/mm) are equal to 38% and 15%, respectively. The detector linearity is very good under the typical mammographic tested exposure conditions
Testing the performances of different image representations for mass classification in digital mammograms
No full textThe classification of tumoral masses and normal breast tissue is targeted. A mass detection algorithm which does not refer explicitly to shape, border, size, contrast or texture of mammographic suspicious regions is evaluated. In the present approach, classification features are embodied by the image representation used to encode suspicious regions. Classification is performed by means of a support vector machine (SVM) classifier. To investigate whether improvements can be achieved with respect to a previously proposed overcomplete wavelet image representation, a pixel and a discrete wavelet image representations are developed and tested. Evaluation is performed by extracting 6000 suspicious regions from the digital database for screening mammography (DDSM) collected by the University of South Florida (USF). More specifically, 1000 regions representing
biopsy-proven tumoral masses (either benign or malignant) and 5000 regions representing normal breast tissue are extracted. Results demonstrate very high performance levels.
The area Az under the receiver operating characteristic (ROC) curve reaches values of 0.973 +/- 0.002, 0.948 +/- 0.004 and 0.956 +/- 0.003 for the pixel, discrete wavelet and overcomplete wavelet image representations, respectively. In particular, the improvement in the Az value with the pixel image representation is statistically significant compared to that obtained with the discrete wavelet and overcomplete wavelet image representations (two-tailed p-value < 0.0001). Additionally, 90% true positive fraction (TPF) values are achieved with false positive fraction (FPF) values of 6%, 11% and 7%, respectively
A ranklet-based CAD for digital mammography
No full textA novel approach to the detection of masses and clustered
microcalcification is presented. Lesion detection is considered as a two-class
pattern recognition problem. In order to get an effective and stable
representation, the detection scheme codifies the image by using a ranklet
transform. The vectors of ranklet coefficients obtained are classified by means
of an SVM classifier. Our approach has two main advantages. First it does not
need any feature selected by the trainer. Second, it is quite stable, with respect
to the image histogram. That allows us to tune the detection parameters in one
database and use the trained CAD on other databases without needing any
adjustment. In this paper, training is accomplished on images coming from
different databases (both digitized and digital). Test results are calculated on
images coming from a few FFDM Giotto Image MD clinical units. The
sensitivity of our CAD system is about 85% with a false-positive rate of 0.5
marks per image
Tomographic approach to single-photon breast cancer imaging with a dedicated dual-head camera with VAOR (SPEMT): Detector characterization
No full textWe have developed a compact Single Photon Emission MammoTomography (SPEMT) scanner capable of imaging the breast for the detection of small size (T1b) tumors. The scanner has a vertical-axis-of-rotation (VAOR) geometry, in which two gamma cameras orbit around a pendulous breast of a prone patient. The SPECT system is rotating around the vicinity of the breast in order to achieve high spatial resolution. The system field-of-view is 147 mm diameter and 41.6 mm height. Each head is made up of one pixilated Nal(TI) crystal matrix with 2.2 mm pitch and 6 mm thickness coupled to three Hamamatsu H8500 64-anodes PMT's. The measured performance confirm that the system could overcome the present clinical sensitivity limit (about 1 cm diameter) for the detection of small size tumors
A Novel Featureless Approach to Mass Detection in Digital Mammograms Based on Support Vector Machines
No full textIn this work, we present a novel approach to mass detection in digital mammograms. The great variability of the masses appearance is the main obstacle of building a mass detection method. It is indeed demanding to characterize all the varieties of masses with a reduced set of features. Hence, in our approach we have chosen not to extract any feature, for the detection of the region of interest; on the contrary, we exploit all the information available on the image. A multiresolution overcomplete wavelet representation is performed, in order to codify the image with redundancy of information. The vectors of the very-large space obtained are then provided to a first SVM classifier. The detection task is here considered as a two-class pattern recognition problem: crops are classified as suspect or not, by using this SVM classifier. False candidates are eliminated with a second cascaded SVM. To further reduce the number of false positives, an ensemble of experts is applied: the final suspect regions are achieved by using a voting strategy. The sensitivity of the presented system is nearly 80% with a false-positive rate of 1.1 marks per image, estimated on images coming from the USF DDSM database
Optimization of the acquisition parameters for a SPET system dedicated to breast imaging
No full textThis work is developed within the framework of a larger project, which aims to develop a multimodal CT-SPET system dedicated to breast imaging. The goal of this paper is to optimize the choice of the various parameters involved in the design of a SPET system dedicated to breast imaging. In particular, we simulated different collimators, different tumor to background (T/B) ratios for two different spherical tumors with diameters of 5 mm and 8 mm. The performance of the explored cameras were analyzed in terms of SNR and image contrast (IC) values, calculated on the reconstructed images. In addition, we investigated the visibility limits of the system, by modifying the tumor size, the T/B value, and the diameter of the breast phantom (8 cm, 10 cm, and 13 cm). As a general tendency, we found out that a high-resolution camera is preferable, in terms of image contrast On the other hand, the General Purpose collimator seems to give a smoother image, giving rise to SNR values comparable to those obtained with the High-Resolution collimator, even with a reduced contrast. High-sensitivity collimators seem to give a worse response on the reconstructed images. The 8 mm tumor is clearly visible for all the simulated conditions, even if it could be very close to the visibility limit for the High-Sensitivity collimator. The 5 mm tumor is close to the visibility limit for General Purpose and High-Resolution collimators, for a T/B ratio equal to 10:1 and is not visible with High-Sensitivity collimator. The smaller tumor is almost obscured by the background with the thickest breast (13 cm diameter)
Testing the performance of image representations for mass classification in digital mammograms
No full textIn this paper a two-class classification problem is faced. One class is constituted
by tumoral masses, breast tumors with size ranging from 3 mm to 30 mm. The
other class is constituted by non-masses. A Support Vector Machine (SVM) is used
as a classifier. Both, masses and non-masses, are extracted from the University
of South Florida (USF) mammographic image database and are presented to the
classifier as crops with pixel size 64 x 64. In order to find the optimal solution to
this problem, different featureless crops representations are evaluated. In particular,
a pixel-based representation, a Discrete Wavelet Transform (DWT) representation
and an Overcomplete Wavelet Transform (OWT) representation are tested
A ranklet-based CAD for digital mammography
No full textAbstract. A novel approach to the detection of masses and clustered microcalcification is presented. Lesion detection is considered as a two-class pattern recognition problem. In order to get an effective and stable representation, the detection scheme codifies the image by using a ranklet transform. The vectors of ranklet coefficients obtained are classified by means of an SVM classifier. Our approach has two main advantages. First it does not need any feature selected by the trainer. Second, it is quite stable, with respect to the image histogram. That allows us to tune the detection parameters in one database and use the trained CAD on other databases without needing any adjustment. In this paper, training is accomplished on images coming from different databases (both digitized and digital). Test results are calculated on images coming from a few FFDM Giotto Image MD clinical units. The sensitivity of our CAD system is about 85 % with a false-positive rate of 0.5 marks per image. 1
