Characterizing mammographic images by using generic texture features.

INTRODUCTION: Although mammographic density is an established risk factor for breast cancer, its use is limited in clinical practice because of a lack of automated and standardized measurement methods. The aims of this study were to evaluate a variety of automated texture features in mammograms as risk factors for breast cancer and to compare them with the percentage mammographic density (PMD) by using a case-control study design. METHODS: A case-control study including 864 cases and 418 controls was analyzed automatically. Four hundred seventy features were explored as possible risk factors for breast cancer. These included statistical features, moment-based features, spectral-energy features, and form-based features. An elaborate variable selection process using logistic regression analyses was performed to identify those features that were associated with case-control status. In addition, PMD was assessed and included in the regression model. RESULTS: Of the 470 image-analysis features explored, 46 remained in the final logistic regression model. An area under the curve of 0.79, with an odds ratio per standard deviation change of 2.88 (95% CI, 2.28 to 3.65), was obtained with validation data. Adding the PMD did not improve the final model. CONCLUSIONS: Using texture features to predict the risk of breast cancer appears feasible. PMD did not show any additional value in this study. With regard to the features assessed, most of the analysis tools appeared to reflect mammographic density, although some features did not correlate with PMD. It remains to be investigated in larger case-control studies whether these features can contribute to increased prediction accuracy.

CADx of mammographic masses and clustered microcalcifications: a review.

Breast cancer is the most common type of cancer among women in the western world. While mammography is regarded as the most effective tool for the detection and diagnosis of breast cancer, the interpretation of mammograms is a difficult and error-prone task. Hence, computer aids have been developed that assist the radiologist in the interpretation of mammograms. Computer-aided detection (CADe) systems address the problem that radiologists often miss signs of cancers that are retrospectively visible in mammograms. Furthermore, computer-aided diagnosis (CADx) systems have been proposed that assist the radiologist in the classification of mammographic lesions as benign or malignant. While a broad variety of approaches to both CADe and CADx systems have been published in the past two decades, an extensive survey of the state of the art is only available for CADe approaches. Therefore, a comprehensive review of the state of the art of CADx approaches is presented in this work. Besides providing a summary, the goals for this article are to identify relations, contradictions, and gaps in literature, and to suggest directions for future research. Because of the vast amount of publications on the topic, this survey is restricted to the two most important types of mammographic lesions: masses and clustered microcalcifications. Furthermore, it focuses on articles published in international journals.

Automatic adaptive enhancement for images obtained with fiberscopic endoscopes.

Modern techniques for medical diagnostics and therapy in keyhole surgery scenarios as well as technical inspection make use of flexible endoscopes. Their characteristic bendable image conductor consists of a very limited number of coated fibers, which leads to so-called comb structure. This effect has a negative impact on further image processing steps such as feature tracking because these overlaid image structures are wrongly detected as image features. With respect to these tasks, we propose an automatic approach to generate optimal spectral filter masks for enhancement of fiberscopic images. We apply the Nyquist-Shannon sampling theorem to the spectrum of fiberscopically acquired images to obtain parameters for optimal filter mask calculation. This can be done automatically and independently of scale and resolution of the image conductor as well as type and resolution of the image sensor. We designed and verified simple rotation invariant masks as well as star-shaped rotation variant masks that contain information about orientation between the fiberscope and sensor. A subjective survey among experts between different modes of filtering certified the best results to the adapted star-shaped mask for high-quality glass fiberscopes. We furthermore define an objective metric to evaluate the results of different filter approaches, which verifies the results of the subjective survey. The proposed approach enables the automated reduction of fiberscopic comb structure. It is adaptive to arbitrary endoscope and sensor combinations. The results give the prospect of a large field of possible applications to reduce fiberscopic structure both for visual optimization in clinical environments and for further digital imaging tasks.

Needs assessment for next generation computer-aided mammography reference image databases and evaluation studies.

INTRODUCTION: Breast cancer is globally a major threat for women's health. Screening and adequate follow-up can significantly reduce the mortality from breast cancer. Human second reading of screening mammograms can increase breast cancer detection rates, whereas this has not been proven for current computer-aided detection systems as "second reader". Critical factors include the detection accuracy of the systems and the screening experience and training of the radiologist with the system. When assessing the performance of systems and system components, the choice of evaluation methods is particularly critical. Core assets herein are reference image databases and statistical methods. METHODS: We have analyzed characteristics and usage of the currently largest publicly available mammography database, the Digital Database for Screening Mammography (DDSM) from the University of South Florida, in literature indexed in Medline, IEEE Xplore, SpringerLink, and SPIE, with respect to type of computer-aided diagnosis (CAD) (detection, CADe, or diagnostics, CADx), selection of database subsets, choice of evaluation method, and quality of descriptions. RESULTS: 59 publications presenting 106 evaluation studies met our selection criteria. In 54 studies (50.9%), the selection of test items (cases, images, regions of interest) extracted from the DDSM was not reproducible. Only 2 CADx studies, not any CADe studies, used the entire DDSM. The number of test items varies from 100 to 6000. Different statistical evaluation methods are chosen. Most common are train/test (34.9% of the studies), leave-one-out (23.6%), and N-fold cross-validation (18.9%). Database-related terminology tends to be imprecise or ambiguous, especially regarding the term "case". DISCUSSION: Overall, both the use of the DDSM as data source for evaluation of mammography CAD systems, and the application of statistical evaluation methods were found highly diverse. Results reported from different studies are therefore hardly comparable. Drawbacks of the DDSM (e.g. varying quality of lesion annotations) may contribute to the reasons. But larger bias seems to be caused by authors' own decisions upon study design. RECOMMENDATIONS/CONCLUSION: For future evaluation studies, we derive a set of 13 recommendations concerning the construction and usage of a test database, as well as the application of statistical evaluation methods.

Detection of malaria parasites in thick blood films.

Malaria, caused by a blood parasite of the genus plasmodium, kills millions of people each year. According to the World Health Organization, the standard for malaria diagnosis is microscopic examination of a stained blood film. We have developed a two-stage algorithm for the automatic detection of plasmodia in thick blood films. The focus of the first stage is on high detection sensitivity while accepting high numbers of false-positive detections per image. The second stage reduces the number of false-positive detections to an acceptable level while maintaining the detection sensitivity of the first stage. The algorithm can detect plasmodia at a sensitivity of 0.97 with a mean number of 0.8 false-positive detections per image. Our results indicate that the proposed algorithm is suitable for the development of an automated microscope for computer-aided malaria screening.

Contour tracing for segmentation of mammographic masses.

CADx systems have the potential to support radiologists in the difficult task of discriminating benign and malignant mammographic lesions. The segmentation of mammographic masses from the background tissue is an important module of CADx systems designed for the characterization of mass lesions. In this work, a novel approach to this task is presented. The segmentation is performed by automatically tracing the mass' contour in-between manually provided landmark points defined on the mass' margin. The performance of the proposed approach is compared to the performance of implementations of three state-of-the-art approaches based on region growing and dynamic programming. For an unbiased comparison of the different segmentation approaches, optimal parameters are selected for each approach by means of tenfold cross-validation and a genetic algorithm. Furthermore, segmentation performance is evaluated on a dataset of ROI and ground-truth pairs. The proposed method outperforms the three state-of-the-art methods. The benchmark dataset will be made available with publication of this paper and will be the first publicly available benchmark dataset for mass segmentation.

Evaluation of spatial interpolation strategies for the removal of comb-structure in fiber-optic images.

Modern techniques for medical diagnosis and therapy in minimal invasive surgery scenarios as well as industrial inspection make considerable use of flexible, fiberoptic endoscopes in order to gain visual access to holes, hollows, antrums and cavities that are difficult to enter and examine. Unfortunately, fiber-optic endoscopes exhibit artifacts in the images that hinder or at worst prevent fundamental image analysis techniques. The dark comb-like artifacts originate from the opaque cladding layer surrounding each single fiber in the image conductor. Although the removal of comb structure is crucial for fiber-optic image analysis, literature covers only a few approaches. Those are based on Fourier analysis and make use of spectral masking or they operate in the spatial domain and rely on interpolation. In this paper, we concentrate on the latter type and introduce interpolation concepts known from related disciplines to the task of comb structure removal. For a quantitative evaluation, we perform experiments with real images as well as with bivariate test functions and rate an algorithm's performance in terms of the normalized root mean square error - a quality metric that it is most commonly used in signal processing for this purpose. Hence, this paper counters the fact that literature lacks an objective performance comparison of the state-of-the-art interpolation based approaches for this type of application.

An evaluation and comparison of the performance of state of the art approaches for the detection of spiculated masses in mammograms.

Mammography is the standard examination method for the early detection of breast cancer. In the last decade, computer assisted detection systems have been developed that assist the physician in the detection of suspicious regions in mammograms. However, recent clinical studies indicate that state of the art CAD systems might have a negative impact on the accuracy of screening mammography. Therefore, besides additional clinical studies, better evaluations of state of the art detection approaches are necessary. In this contribution three methods for the detection of spiculated masses in mammograms are evaluated and compared. All three of them are based on gradient orientation images. To detect masses, the methods use circular neighbourhoods with different sizes around a single pixel. The number of orientations in every neighbourhood is used by every method in different ways to form a result. The main contribution is the first fair comparison of the performance of different detection approaches for spiculated masses. Furthermore, a novel gradient direction analysis is introduced. The analysis is an extension to the three approaches, which increases the performance for one of the three approaches.

Improving the accuracy of feature extraction for flexible endoscope calibration by spatial super resolution.

Many applications in the domain of medical as well as industrial image processing make considerable use of flexible endoscopes - so called fiberscopes - to gain visual access to holes, hollows, antrums and cavities that are difficult to enter and examine. For a complete exploration and understanding of an antrum, 3d depth information might be desirable or yet necessary. This often requires the mapping of 3d world coordinates to 2d image coordinates which is estimated by camera calibration. In order to retrieve useful results, the precise extraction of the imaged calibration pattern's markers plays a decisive role in the camera calibration process. Unfortunately, when utilizing fiberscopes, the image conductor introduces a disturbing comb structure to the images that anticipates a (precise) marker extraction. Since the calibration quality crucially depends on subpixel-precise calibration marker positions, we apply static comb structure removal algorithms along with a dynamic spatial resolution enhancement method in order to improve the feature extraction accuracy. In our experiments, we demonstrate that our approach results in a more accurate calibration of flexible endoscopes and thus allows for a more precise reconstruction of 3d information from fiberoptic images.

Physically motivated enhancement of color images for fiber endoscopy.

Fiber optics are widely used in flexible endoscopes which are indispensable for many applications in diagnosis and therapy. Computer-aided use of fiberscopes requires a digital sensor mounted at the proximal end. Most commercially available cameras for endoscopy provide the images by means of a regular grid of color filters what is known as the Bayer Pattern. Hence, the images suffer from false colored spatial moiré, which is further stressed by the downgrading fiber optic transmission yielding a honey comb pattern. To solve this problem we propose a new approach that extends the interpolation between known intensities of registered fibers to multi channel color applications. The inventive idea takes into account both the Gaussian intensity distribution of each fiber and the physical color distribution of the Bayer pattern. Individual color factors for interpolation of each fiber area make it possible to simultaneously remove both the comb structure from the fiber bundle as well as the Bayer pattern mosaicking from the sensor while preserving depicted structures and textures in the scene.