Yongbum Lee

Nodule detection on chest helical CT scans by using a genetic algorithm

The purpose of the study is to apply a genetic algorithm (GA) template matching method to detect lung nodules in chest helical X ray CT (computed tomography) images. We combined GA and template matching to search the positions of nodules and to calculate adaptation scales of individuals on GA, respectively. We used four simulated nodules created by Gaussian distribution, whose sizes were different to each other, as reference patterns in the GA template matching. The GA selected an adequate reference image from four images and searched adequate positions to template matching. We used cross correlation as similarity of template matching and as adaptation scales of individuals on GA. It was possible to detect 23 nodules from 45 that did not touch the lung walls, without consideration of their sizes. It was also possible to detect all nodules that touched the lung walls by using conventional template matching along lung walls. The total detection rate was approximately 67%. The number of false positives per slice was over 10. To improve the detection performance and to decrease the number of false positives, we are now working on considering operators and their parameters of GA.

Medical image classification using genetic-algorithm based fuzzy-logic approach

In this paper we present a genetic-algorithm-based fuzzy-logic approach for computer-aided diagnosis scheme in medical imaging. The scheme is applied to discriminate myocardial heart disease from echocardiographic images and to detect and classify clustered microcalcifications from mammograms. Unlike the conventional types of membership functions such as trapezoid, triangle, S curve, and singleton used in fuzzy reasoning, Gaussian-distributed fuzzy membership functions (GDMFs) are employed in the present study. The GDMFs are initially generated using various texture-based features obtained from reference images. Subsequently the shapes of GDMFs are optimized by a genetic-algorithm learning process. After optimization, the classifier is used for disease discrimination. The results of our experiments are very promising. We achieve an average accuracy of 96% for myocardial heart disease and accuracy of 88.5% at 100% sensitivity level for microcalcification on mammograms. The results demonstrated that our proposed genetic-algorithm-based fuzzy-logic approach is an effective method for computer-aided diagnosis in disease classification.

A method of medical image enhancement using wavelet analysis

The paper presents a wavelet-based method for medical image enhancement. A transfer function is derived for transferring wavelet coefficients at various multiresolution levels with different weighting values. Applications of the method to digital radiographic images of chest X-rays are shown. Moreover, in order to validate the effectiveness of the proposed method, we compare the results obtained using the wavelet analysis method to that using the conventional fast Fourier transform (FFT) method. The results show that the proposed method is superior to the FFT method in terms of image enhancement.

A method of medical image enhancement using wavelet-coefficient mapping functions

An approach to medical image enhancement is proposed. In this method, non-linear mapping functions are derived for projecting a set of discrete wavelet transform (DWT) coefficients to a new set of DWT coefficients. The mapped coefficients are then inverse wavelet transformed. Applications of the proposed method to the chest x-ray images and mammograms with subtle micro-calcifications are shown. To validate the effectiveness of the proposed method, we compared the results to those obtained by the fast Fourier transform (FFT) and the conventional wavelet-based (CWB) methods. Our preliminary results strongly suggest that the proposed method offers considerably improved enhancement capability over the FFT and CWB methods.

A Modified Undecimated Discrete Wavelet Transform Based Approach to Mammographic Image Denoising

In this work, the authors present an effective denoising method to attempt reducing the noise in mammographic images. The method is based on using hierarchical correlation of the coefficients of discrete stationary wavelet transforms. The features of the proposed technique include iterative use of undecimated multi-directional wavelet transforms at adjacent scales. To validate the proposed method, computer simulations were conducted, followed by its applications to clinical mammograms. Mutual information originating from information theory was used as an evaluation measure for selection of an optimal wavelet basis function. We examined the performance of the proposed method by comparing it with the conventional undecimated discrete wavelet transform (UDWT) method in terms of processing time-consuming and image quality. Our results showed that with the use of the proposed method the computation time can be reduced to approximately 1/10 of the conventional UDWT method consumed. The results of visual assessment indicated that the images processed with the proposed UDWT method showed statistically significant superior image quality over those processed with the conventional UDWT method. Our research results demonstrate the superiority and effectiveness of the proposed approach.

Z-score Mapping Method for Extracting Hypoattenuation Areas of Hyperacute Stroke in Unenhanced CT

RATIONALE AND OBJECTIVES The purpose of this study was to develop a z-score mapping method on the basis of a voxel-by-voxel analysis to visualize hypoattenuation areas of hyperacute stroke on unenhanced computed tomographic (CT) images. MATERIALS AND METHODS The algorithm of the developed method consisted of five main steps: anatomic standardization, the construction of a normal reference database, calculation of the z scores, the elimination of false-positive areas, and the extraction of hypoattenuation areas. The obtained z-score map was then superimposed on the original CT images for identifying hypoattenuation areas of hyperacute stroke on the unenhanced CT images. The method was applied to 21 patients with infarctions of the middle cerebral artery territory <3 hours after symptom onset. The performance of the method was evaluated using receiver-operating characteristic analysis. RESULTS Hypoattenuation regions could be significantly distinguished from normal regions by z-score values (P < .0001). The area under the receiver-operating characteristic curve for distinction between 68 hypoattenuation regions and 142 normal regions was 0.834. CONCLUSIONS The developed method has the potential to accurately indicate high-signal intensity areas corresponding to hypoattenuation areas on CT images in the hyperacute stage of stroke.

Automatic determination of the imaging plane in lumbar MRI

In this paper we describe a method for assisting radiological technologists in their routine work to automatically determine the imaging plane in lumbar MRI. The method is first to recognize the spinal cord and the intervertebral disk (ID) from the lumbar vertebra 3-plane localizer image, and then the imaging plane is automatically determined according to the recognition results. To determine the imaging plane, the spinal cord and the ID are automatically recognized from the lumbar vertebra 3-plane localizer image with a series of image processing techniques. The proposed method consists of three major steps. First, after removing the air and fat regions from the 3-plane localizer image by use of histogram analysis, the rachis region is specified with Sobel edge detection filter. Second, the spinal cord and the ID were respectively extracted from the specified rachis region making use of global thresholding and the line detection filter. Finally, the imaging plane is determined by finding the straight line between the spinal cord and the ID with the Hough transform. Image data of 10 healthy volunteers were used for investigation. To validate the usefulness of our proposed method, manual determination of the imaging plane was also conducted by five experienced radiological technologists. Our experimental results showed that the concordance rate between the manual setting and automatic determination reached to 90%. Moreover, a remarkable reduction in execution time for imaging-plane determination was also achieved.

Detectability improvement of early sign of acute stroke on brain CT images using an adaptive partial smoothing filter

Detection of early infarct signs on non-enhanced CT is mandatory in patients with acute ischemic stroke. We present a method for improving the detectability of early infarct signs of acute ischemic stroke. This approach is considered as the first step for computer-aided diagnosis in acute ischemic stroke. Obscuration of the gray-white matter interface at the lentiform nucleus or the insular ribbon has been an important early infarct sign, which affects decisions on thrombolytic therapy. However, its detection is difficult, since the early infarct sign is subtle hypoattenuation. In order to improve the detectability of the early infarct sign, an image processing being able to reduce local noise with edges preserved is desirable. To cope with this issue, we devised an adaptive partial smoothing filter (APSF). Because the APSF can markedly improve the visibility of the normal gray-white matter interface, the detection of conspicuity of obscuration of gray-white matter interface due to hypoattenuation could be increased. The APSF is a specifically designed filter used to perform local smoothing using a variable filter size determined by the distribution of pixel values of edges in the region of interest. By adjusting four parameters of the APSF, an optimal condition for image enhancement can be obtained. In order to determine a major one of the parameters, preliminary simulation was performed by using composite images simulated the gray-white matter. The APSF based on preliminary simulation was applied to several clinical CT scans in hyperacute stroke patients. The results showed that the detectability of early infarct signs is much improved.

Automated lesion detection methods for 2D and 3D chest X-ray images

The purpose of this work is to present some technical approaches of our computer-aided detection (CAD) system for chest radiograms and helical CT scans, and also evaluate that by using three databases. The CAD includes some methods to detect lesions and to eliminate false-positive findings. The detection methods consist of template matching and artificial neural network approaches. A genetic algorithm (GA) was employed in template matching to select a matched image from various reference patterns. Artificial neural networks (ANN) were also applied to eliminate the false-positive candidates. The sensitivity and the number of false-positives were 73% and 11 FP per image on chest radiogram CAD and 77% with 2.6 FP per image on helical CT scan CAD. These preliminary results imply that the GA and ANN-based detection methods may be effective in indicating lesions on chest radiograms and helical CT scans.

Improvement in automated detection of pulmonary nodules on helical x-ray CT images

We previously developed a scheme to automatically detect pulmonary nodules on CT images, as a part of computer-aided diagnosis (CAD) system. The proposed method consisted of two template-matching approaches based on simple models that simulate real nodules. One was a new template-matching technique based on a genetic algorithm (GA) template matching (GATM) for detecting nodules within the lung area. The other one was a conventional template matching along the lung wall [lung wall template matching (LWTM)] for detecting nodules on the lung wall. After the two template matchings, thirteen feature values were calculated and used for eliminating false positives. Twenty clinical cases involving a total of 557 sectional images were applied; 71 nodules out of 98 were correctly detected with the number of false positives at approximately 30.8/case by applying two template matchings (GATM and LWTM) and elimination process of false positives. In this study, five features were newly added, and threshold-values of our previous features were reconsidered for further eliminating false positives. As the result, the number of false positives was decreased to 5.5/case without elimination of true positives.