This paper describes the technology for building virtual environments, which realize substantial and ractical medical diagnostic functions. The authors introduce both passive and active visualizations. While passive visualization is just an observation of organs' shape and function, active visualization brings "realization" system, in which users can immerse into the diseased part, execute quantitative measurement of characteristic parameters, and reflect users' intention to visualization.

View full abstract

In this paper, we describe an automatic extraction method of the bone axes of femur and tibia using inferior limb image though it needs leg information on a left leg or a right leg beforehand. Moreover, our image processing method is applicable for virtual CR image generated by CT image. Since the method can be applied for both front and side virtual CR images from CT image, accurate 3D coordinates of 3D bone axis are calculated efficiently. The bone axis information is effective for labor saving measure for doctors, precise surgery planning, and useful for 2D and 3D implant template placements. Moreover, since 3D bone axis is precisely defined in 3D space, it is possible to place bone cutting planes on orthogonal line of bone axis, to decide rotation angles of implants, and to compare each loading axis and bone axis.

View full abstract

Analyzing 3-D knee kinematics of implanted knee after total knee arthroplasty (TKA) plays a key role of TKA research and diagnosis. This article introduces an automated method using 2-D digital radiography (DR) video. Although there are some conventional studies based on 2-D/3-D image registration of 2-D DR image and 3-D geometric model, they are designed for still image analysis because the image analysis requires human interaction for initializing 3-D pose/position parameter in 6 degree-of-freedom (DOF) parameter search space. The present method introduces Monte Calro method for initializing and searching 3-D pose/position parameter. Thus, it can analyze all frame of DR video with no user interaction. This method was applied to 7 subjects, and some kinematics patterns were shown.

View full abstract

We propose a novel intersection profile method for a four-dimensional magnetic resonance image (4DMRI) to analyze thoracic motion from time-sequential images of two-dimensional (2D) MR images acquired under free breathing. Using the proposed method, first, time-sequential MR images in many coronal planes that are set to cover the lung area widely are acquired as data slices. Second, the timesequential MR images in a proper sagittal plane are acquired as navigator slices. Then 2D spatio-temporal images (2DSTI), constituting a time-sequential array of profiles along the intersection of slices of two kinds, are generated for both the navigator slice and the data slices. After a proper respiratory pattern (template) is extracted from 2DSTI of the navigator slice, template matching by normalized cross-correlation is conducted to identify a similar respiratory pattern from the 2DSTI of the data slices. Then 4DMRI is reconstructed by application of this process to each data slice. Additionally, to evaluate diaphragmatic motion quantitatively, we constructed a diaphragmatic function map using 4DMRI reconstructed using the proposed method. The diaphragmatic function map was able to visualize diaphragmatic function of spontaneous breathing.

View full abstract

The image registration of a set of images obtained at different stages of diagnosis, treatment, and evaluation is important for assessment of the therapeutic effect and changes in lesions over time in order to determine an effective treatment policy. Image-registration studies have been conducted in the fields of CT, MR, and PET. However, few studies have been reported on the registration of ultrasound images, although it is the most widely used diagnostic imaging devices for circulatory system and abdomen. In recent years, 4D ultrasound imaging systems have been developed that observe and record 3D volume data by scanning 2D ultrasound images electrically or mechanically. 4D systems are strongly expected to facilitate the registration of ultrasound images obtained at the various stages of the diagnostic processes. In this paper, we introduce registration of pre-operative, intra-operative, and post-operative images with the help of an improved registration method, surgical assistance by sequential registration of 4D volume data during surgery, and the development of 3D panoramic volume by the automatic registration of successive 3D volume data.

View full abstract

We have been researching on Global Collaboration Medical Network System, in which we consider the network infrastructure as the basic functions for the regional medical collaboration between local hospitals and clinics. In this paper, we propose and implement 3D Electronic Referral Letters system. We add the voice and multi dimensional movie support functions on the original system. This system will help to improve the medical image applications such as regional medical collaboration, remote medicine, emergent medicine and home medical care.

View full abstract

An ultra-short TE pulse sequence is composed of a radial sampling that acquires echo signals radially in the K-space and a half-echo acquisition that acquires only half of the echo signal. The shift in the position of the echo signal (echo shift) caused by the timing errors in the gradient magnetic field pulses affects the image quality in the radial sampling with the half-echo acquisition. To improve image quality, we have developed a signal correction algorithm that detects and eliminates this echo shift during reconstruction by performing a pre-scan within 10 seconds. The results showed that image quality is improved under oblique and/or off-centering conditions that frequently cause image distortion due to hardware error. In conclusion, we have developed a robust ultra-short TE pulse sequence that allows wide latitude in the scan parameters, including oblique and off-centering conditions.

View full abstract

Contralateral subtraction (C-sub) is one of a number of computer-aided diagnostic techniques for detecting asymmetric abnormalities in chest radiographs. In this technique, pulmonary nodules in chest radiographs can be enhanced by subtracting a left/right reversed mirror image from the original image. In this report, we propose a C-sub technique that employs novel global and local registration methods. We evaluated the subtraction images obtained by applying the proposed C-sub technique to 107 images with nodules in the JSRT database for which the subtlety levels were judged to be "very subtle", "subtle", and "relatively obvious". First, the quality of rib elimination in the subtraction images was evaluated by a radiologist and assigned a point score (3 = acceptable or better, 2 = poor, and 1 = very poor). In this evaluation, 92.5% of cases were scored as 3. Next, whether or not the nodules were clearly depicted at the specified locations in the subtraction images was evaluated by a radiologist, and the quality of nodule depiction was assigned a point score (3 = clearly depicted, 2 = faintly depicted, and 1 = not depicted). In this evaluation, 73.8% of cases were scored as 3 and 7.5% of cases were scored as 2. The time needed to apply the proposed C-sub technique to the 107 images evaluated was 19.4 seconds per image on average using a personal computer with a 2.2-GHz Intel processor.

View full abstract

In this report, we propose a method for denoising the tissue time-activity curve (tTAC) obtained from a series of brain PET images. The tTAC is used in a various types of receptor analysis such as a linear method for estimating the total volume of distribution (V_T). However, the presence of a large amount of noise in PET scanning reduces the accuracy of the estimation. In the proposed method, the measured tTAC is denoised using the parametric eigenspace of the tTAC. It is known that the tTAC is well represented by the compartment model, which is derived from the behavior of the ligand in the brain. In order to eliminate noise, we compute the eigenspace of a set of tTACs that have been artificially generated based on the compartment model. Given an observed tTAC, we project it to the eigenspace and eliminate the noise by means of MAP estimation. We evaluated the denoising performance of the proposed method using both simulated tTACs and actual clinical data. The V_T was estimated using the denoised tTACs. The evaluation results showed that the bias and variance of the estimates were improved by employing our proposed denoising method. This indicates that the proposed method is useful for denoising tTACs.

View full abstract

In this report, we present an image evaluation method in which mutual information (MI) is employed as a metric for selecting the optimal wavelet basis function to be used in denoising planar nuclear images, with the objective of improving image quality. The higher the MI value, the better the image quality. We initially selected eight different wavelet basis functions for investigation in the present study. Subsequently, wavelet transforms were applied to planar images for denoising by employing the universal soft-thresholding method. Finally, the MI values of the wavelet-transformed images were computed for comparison. In this study, a computer-generated 2-D grid-pattern image and phantom images produced using a standard inkjet printer served as the original images. The results for the simulation and phantom images showed the same trend of ranking in terms of MI. The images processed by dmey wavelet showed the highest MI values. To validate the usefulness of the proposed method, the standard deviation rate and edge slope ratio of the processed images were calculated and compared. The results showed that the MI value can serve as an effective criterion for selecting the optimal wavelet basis function for image denoising. The results also showed that, of the eight wavelet basis functions investigated, dmey wavelet is the optimal basis function for denoising low-contrast planar images.

View full abstract

[in Japanese]

View full abstract