Thor Ole Gulsrud

Watershed segmentation of detected masses in digital mammograms

A method for segmentation of detected masses in digital mammograms is introduced. The method is based on gray scale mathematical morphology. In a preprocessing step, image enhancement based on a local histogram technique is applied, followed by a morphological smoothing operation. The watershed transform is then applied to the gradient of the smoothed image resulting in segmented regions. A good segmentation is important in order to be able to extract useful feature measures from the segmented regions. These feature measures can be input to a classifier which classifies each region as either a mass or a false detection. Initial experiments have been performed using mammograms from the MIAS database. Results of the experimental study indicate that our scheme can provide useful contour extraction for mass structures

A filter-based approach towards automatic detection of microcalcification

To establish a practical CAD (Computer-Aided Diagnosis) system to facilitate the diagnosis of microcalcifications, we propose a filter-based technique to detect microcalcifications. Via examination of an existing optimal filter-based technique, it is found that its performance on highlighting the energy of mammograms is seriously affected by artefacts and the background of breast. As a result, four methods in pre and post-processing are described in this paper to improve the optimal filtering, leading to an adaptive selection of thresholds for input mammograms. These methods have been tested by using 30 mammograms (with 25 microcalcifications) from the MIAS database and 23 mammograms from DDSM database. Comparing with the original optimal filter-based technique, our technique reduces the false detections (FD), eliminates the influence of the background in mammograms and is able to adaptively select the threshold for the detection of microcalcifications.

Image texture classification with digital filter banks and transforms

Several frequency domain or joint spatial/frequency domain techniques for image texture classification have been published. We formulate these techniques within a common signal processing framework based on digital filter banks. The usefulness of computationally efficient IIR filter banks as channel filters in texture classifiers is demonstrated. Using estimates of local energy in the frequency channels we also propose a technique for selecting optimum filter banks by maximizing a between class distance measure. This optimization is particularly simple when using the IIR based filter banks.

Analysis of mammographic microcalcifications using a computationally efficient filter bank

We present a new method for classification of malignant and benign clusters of microcalcifications in digital mammograms. A computationally efficient infinite impulse response (IIR) quadrature mirror filter (QMF) bank is used as a tool for extracting texture and shape features. The filter bank splits the input image into four subbands: low-low band, low-high band, high-low band and high-high band. Texture and shape features based on cooccurrence matrices are computed from the subsampled subbands. The low-low band extracts the information of spatial dependence and the higher frequency bands extract the shape information. The results of an experimental study demonstrate that our approach drastically improves the overall performance compared to a manual system.

Detection of circumscribed masses in mammograms using morphological segmentation

We present a method for detecting circumscribed masses in digital mammograms. Morphological hierarchical watersheds are used in the segmentation process. Oversegmentation is prevented by employing a reconstructive open/close alternating sequential filter to simplify the image. The advantage of this method of simplification is that the object shapes and edges are preserved. The regional maxima of the simplified input image are then extracted and used as internal markers for the hierarchical watershed transform, which is applied to the gradient image of the simplified input image. An image-based classification technique is applied to reduce the number of false positives. The method is applied to 18 mammograms from the MIAS database, containing 20 circumscribed masses in background tissue of varying density. We obtain a high true detection rate using combined with a low number of false positives per image.

Watermarking of digital mammograms without interfering with automatic detection of microcalcifications

In this paper we show that it is possible to watermark digital mammograms with patient information without interfering significantly with an existing method for automatic detection of microcalcifications. The paper discuss which type of information that should be embedded in digital medical data, i.e. what the digital watermark should consists of. Robustness requirements relevant for this watermarking application are discussed and taken into account.

Classification of microcalcifications using a multichannel filtering approach

Presents a method for classification of Mammographic Microcalcifications using Quadrature Mirror Filter (QMF) bank subband decomposition as a tool for extracting textural features. The QMF bank splits the input image into four subbands: low-low band, low-high band, high-low band and high-high band, and textural features based on co-occurrence matrices are computed from the subsampled subbands. The low-low band extracts the information of spatial dependence and the higher frequency bands extract the structural information. The experiments demonstrate that our approach can provide significant discrimination for classification of benign and malignant microcalcifications.

Image texture classification using Quadrature Mirror Filter bank in combination with Co—occurrence matrices

This paper presents a method for the classification of textures using quadrature mirror filter (QMF) bank subband decomposition in combination with statistical descriptors. In our combined method the QMF bank splits the input image into four subbands, and statistical descriptors based on co-occurrence matrices are computed from the subsampled low-low band. The experiments demonstrate that the combined method has better classification performance than that of statistical descriptors computed from the co-occurrence matrices of the whole texture image. In addition, the experiments demonstrate that the combined method based on computationally efficient IIR QMF banks yields approximately the same classification results as the combined method based on classical FIR QMF banks.

Feature extraction in digital mammograms based on optimal and morphological filtering

The aim of this study is to provide feature images of digital mammograms in which regions corresponding to masses are enhanced. Subsequently, the feature images can be segmented and classified into two classes; masses and normal tissue. Our proposed feature extraction method is based on a local energy measure as texture feature. The local energy measure is extracted using a filter optimized with respect to the relative distance between the average feature values. In order to increase the sensitivity of the texture feature extraction scheme each mammogram is preprocessed using wavelet transformation, adaptive histogram equalization, and a morphology based enhancement technique. Initial experiments indicate that our scheme is able to provide useful feature images of digital mammograms. In order to quantify the system performance the feature images of 38 mammograms from the MIAS database -- 19 containing circumscribed masses, and 19 containing spiculated masses -- were segmented using simple gray level thresholding. For the circumscribed masses a true positive (TP) rate of 89% with a corresponding 2.3 false detections (false positives, FPs) per image was achieved. For the spiculated masses the performance was somewhat lower, yielding an overall TP rate of 84% with a corresponding 2.6 FPs per image.

Compression of digital mammograms with region-of-interest coding evaluated on a CAD system

Screening programs produce large amount of mammographic data, and good compression schemes would be beneficial for both storage and transmission purposes. In medical data it is crucial that diagnostic important information is preserved. In this work we have implemented two different region-of-interest (ROI) coding methods together with a Set Partitioning in Hierarchical Trees (SPIHT) scheme to be used for compression of mammograms. Region-of-interest coding allows a region of the image to be compressed with higher fidelity than the rest of the image. This is useful in medical data to be able to compress a region containing a possibly cancer area with very high fidelity, but still manage an overall good compression ratio. Both the ROI methods, the basic SPIHT method as well as JPEG compression standard, the latter two without possibility of ROI coding, are evaluated by studying the results from a Computer Aided Detection (CAD) system for microcalcifications tested on the original and the compressed mammograms. In addition a visual inspection is performed as well as Peak Signal-to-Noise-Ratio (PSNR) calculations. Mammograms from the MIAS database is used. We show that mammograms can be compressed to less than 0.5 (0.3) bpp without any visual degradation and without significantly influence on the performance of the CAD system.