Search PubMed⌕ Search

Biomedical subjects

Y N Sun

Publications and source records attributed to Y N Sun.

30 records · Page 2Linked to original sources
PubMed1994

A computer system for skeletal growth measurement.

In this paper, we have presented a new image-processing system for the measurement of skeletal growth in pediatric radiology. From a standard posterior and anterior view radiograph, taken from a left hand, the proposed system first automatically locates the phalangeal region of interest, and then measures the geometrical parameters associated with skeletal maturity. Finally, the bone age is estimated by using the standard phalangeal length table. Clinical studies reveal that the computer processing has resulted in an objective and accurate assessment of skeletal age. It greatly improves the shortcomings, including inter- and intraobserver variations and inaccuracy, reported in other research by manual methods. In conclusion, it is an inexpensive and useful tool for the evaluation of short-term abnormalities in the skeletal growth of children.

Adolescent↗
PubMed1994

Measurement of organ volume by single photon emission computed tomography.

The aim of this study was to develop a new algorithm for volume determination by single photon emission computed tomography (SPECT). Different algorithms were evaluated through phantom studies. The results show that the algorithm combining moment-preserving bilevel thresholding and best-fit Laplacian second derivative edge detection can provide the most accurate measurement of volume. Besides, this method can be utilized in different SPECT systems with no need for further phantom studies. In patient studies, the results of liver volume calculation have indicated that this is a useful technique in clinical medicine.

Algorithms↗
PubMed1993

Two E2 binding sites (E2BS) alone or one E2BS plus an A/T-rich region are minimal requirements for the replication of the human papillomavirus type 11 origin.

Human papillomaviruses (HPVs) cannot be propagated in vitro, but the DNA can be replicated transiently in an assay in the presence of two trans-acting viral proteins, E1 and E2. Using this assay, we have defined the minimal cis-acting elements of the origin of replication of HPV type 11. Most HPV genomes are conserved at the origin of replication, and the core contains three E2 binding sites (E2BS) surrounding an A/T-rich spacer region. The present results show that the minimal requirement for replication is either two E2BS alone or the A/T-rich region plus one E2BS; in the latter case the relative position of the E2BS is important. In all the studies, the presence of both E1 and E2 proteins was essential for replication, yet only the E2BS was required at the origin. We have shown that E1, E2, and the origin of replication containing an E2BS from a complex in vitro, and our data are consistent with a model in which E2 acts to target E1 to the HPV type 11 replication origin.

Base Sequence↗
PubMed1993

[Ectopic HCG in gastric carcinoma detected by immunohistochemical technique].

Eighty-eight cases of gastric carcinoma were tested for ectopic production of HCG with anti-HCG antibody by PAP method. Among 88 cases of gastric carcinoma, 17 cases were positive for HCG. Tumors were classified as well or moderately or poorly differentiated. In the poorly differentiated group, 53.85% of the tumors were immunopositive for HCG, and 66.67% had metastasized to lymph nodes. It is suggested that gastric carcinoma may produce HCG, and HCG-positive tumors tend to have higher malignancy and poor histological differentiation. The possible origin and histogenesis of HCG-producing tumor cells are discussed.

Adenocarcinoma↗
PubMedDate not supplied

A self-learning segmentation framework--the Taguchi approach.

The detection of object boundary is an interesting and challenging task in computer vision and medical image processing. The active contour model (snake model) has attracted much attention for object boundary detection in the past decade. However, due to the lack of understanding on the effect of different energy terms to the behavior of related objective functions for an image, the assignment of weights for different energy terms in this model is usually fulfilled empirically. Few discussions have been brought out specifically for assigning these weights automatically. In this paper, a novel self-learning segmentation framework, based on the snake model is proposed and applied to the detection of cardiac boundaries from ultrasonic images. The framework consists of a learning section and a detection section, and provides a training mechanism to obtain the weights from a desired object contour given manually. This mechanism first employs Taguchi's method to determine the weight ratios among distinct energy terms, followed by a weight refinement step with a genetic algorithm. The refined weights can be treated as the a priori knowledge embedded in the manually defined contour and be used for subsequent contour detection. Experiments with both synthetic and real echocardiac images were conducted with satisfactory outcomes. Results also show that the present method can be used to analyze successive images of the same object with only one training contour. Finally, the validity of the weight determining process was verified by the analysis of variance method (ANOVA).

Algorithms↗
PubMedDate not supplied

Voxel-based texture mapping for medical data.

In computerized image and graphic applications, texture mapping is one of the most commonly used methods to improve the realism or to enhance the visual effect of object rendering without too much increase in computational complexity. The conventional method usually has to transfer three-dimensional (3D) object to the polygonal structure, and is computationally expensive. As the medical data are mostly in voxel format, the polygonal structure is not efficient or requires more complicated mechanism in retrieving the internal information of medical data. In this paper, we propose a new texture mapping method, based on flattening a chain-coded 3D surface, to handle the voxel-based data directly. The method flattens the 3D object surface onto a two-dimensional (2D) plane and then uses 2D metamorphosis to generate the correspondences between object surface and texture image. Therefore, polygon transformation is no longer necessary and texture mapping is handled with inexpensive 2D morphing. More importantly, the internal information of medical data can be easily preserved and utilized further. Experimental results have shown the effectiveness and efficiency of the proposed algorithm.

Algorithms↗
PubMedDate not supplied

Edge-shrinking interpolation for medical images.

A new algorithm for interpolating the missing data between two adjacent medical images is presented. Our method is useful for solving the interpolation of any region-represented images of an object to be reconstructed, even when the object is stretched abruptly, branched or hollow, as often occurs in medical images, which cases can not be handled well by existing methods. When this algorithm is applied, the nonoverlapped regions of the same object in the two base images are first extracted and encoded by chamfer distance code on every pixel in these regions. Then, the outer edges of the nonoverlapping regions are shrunk inward simultaneously so that the stretched edges reach the edges of the overlapping regions at the same time. The distance codes in nonoverlapping regions are used to limit the shrinking of these edges in the interpolation process. The proposed method also provides object centralization and enlargement operations to obtain stable and reasonable results in complicated case. The experimental results show that the proposed method is more efficient in resolving general interpolation tasks than the existing methods (S.P. Raya and J. K. Udupa, IEEE Trans. Med. Image. 9, 32-42, 1990; G. T. Herman et al., IEEE Comput. Graph. Appl. 12, 69-79, 1992; J.F. Guo et al., Comput. Med. Imag. Graph. 19, 267-279, 1995.)

Algorithms↗
PubMedDate not supplied

Interactive voxel surface rendering in medical applications.

Semi-boundary (SB) data structure is a compact voxel surface representation of the structure from the medical images. It represents only the boundary of the extracted structure and only an opaque object boundary involved in a 3D dataset can be visualized. Its computational complexity is in proportion to the number of SB voxels. In this paper, we propose schemes to reduce the number of projections in two ways. First, in conjunction with neighboring code, we exploit a set of visibility tables to cull some of the invisible SB voxels. Second, we exploit three pass rotations and an incremental approach to quickly determine the projection position for each SB voxel during rendering. With these two combinations, we significantly improve SB rendering performance. As a result, we can achieve an interactive rendering speed on general purpose workstations for our medical applications.

Algorithms↗
PubMedDate not supplied

Liver, spleen and tumor volume measured by personal computer.

BACKGROUND/AIMS: Computed tomography (CT) scans are common examinations for patients with chronic liver diseases. To quantitate the organ or tumor volume from the scans and to accomplish the task in an efficient way with the most economic equipment, we developed a system based on a personal computer. METHODOLOGY: We used color-markers and transparency to sketch the edges of liver, hepatoma, and spleen. Each organ or tumor of interest is marked out by fine-point markers on pieces of transparency. The sketch was scanned into a digitized image format on a personal computer (Pentium 133). The calculation involves edge detection, three-dimensional reconstruction, and voxel counting. By using summation-of-the-area and trapezoid approximation technique, the voxels of each structure are counted. In this study, we illustrate the potential application in the management of a hepatic cancer patient. RESULTS: After digitalization, the data size of CT images is about 1 to 1.5 megabytes. It takes less than 5 min to complete volume calculation. CONCLUSIONS: By this method, tumor load before and after chemotherapy can be estimated easily and accurately. This would be helpful in clinical practice.

Carcinoma, Hepatocellular↗
PubMedDate not supplied

Liver volume in patients with or without chronic liver diseases.

BACKGROUND/AIMS: The size of the liver is an important clinical parameter; the aim of this study is to examine the correlation between liver volume and etiology and the severity of disease, and to evaluate its usefulness in predicting survival. METHODOLOGY: Patients observed in this study were comprised of thirty three patients with non-liver disease and 44 patients with chronic liver disease (alcoholic hepatitis, 9; hepatitis B, 24; and hepatitis C, 11). The liver volume was measured from digitized CT scan images. Techniques of planimetry and summation of areas were utilized for calculation. RESULTS: The prediction model to estimate liver volume in patients without liver disease was: liver volume (ml)= [13 x height (cm)] +[12 x weight (Kg)] - 1530. The volume ratio (%) [(volume from reconstructed image /predicted volume) x 100] of alcoholic patients was 135.9+/-25.8, which was significantly higher than that of chronic hepatitis B (73.6+/-15.4) and chronic hepatitis C (74.5+/-20.7). Patients with chronic viral hepatitis were classified into Child-Pugh class A (N=10), B (N=14) and C (N=11). Analysis of variance and trend test revealed that the volume ratio had a significant decreasing trend from the control group (100.5+/-8.1), class A (83.4+/-13.9), class B (72.2+/-13.2) to class C (63.3+/-14.4). CONCLUSIONS: Liver volume can be predicted from patients' weight and height if they have no liver disease. The liver volume ratio correlates much better with etiology and severity of the disease and is a reliable predictor for patient's survival.

Adult↗
PubMedDate not supplied

Blood supply of esophageal stumps.

BACKGROUND/AIMS: The purpose of the study was to determine the blood supply of stumps of mobilized esophagus. METHODOLOGY: Esophageal stumps of five groups of rabbits were studied. The esophageal stumps were dyed via intravenous injection of trypan blue, and were analyzed by color image processing. RESULTS: The dyed length of the esophageal stump was 7.15+/-0.5 cm in the stomach-connecting group; 4.6+/-0.69 cm in the stomach-connecting with myotomy group; 6.1+/-0.88 cm in the distal stomach-connecting group; 3.0+/-0.47 cm in the neck-connecting group; and, 2.1+/-0.2 cm in the neck-connecting with myotomy group. There was a significant difference in the dyed length between the stomach-connecting and neck-connecting groups (p=0.000), and between the stomach-connecting and stomach-connecting with myotomy groups (p=0.026), but there was no significant difference between the neck-connecting and neck-connecting with myotomy (p=0.094), or the stomach-connecting and distal stomach-connecting groups (p=0.053). CONCLUSIONS: Stumps of the mobilized esophagus in rabbits, with or without myotomy, exhibit differences in dyed length, depending on the diameter, wall thickness and location of the blood supply in the esophagus.

Anastomosis, Surgical↗