• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.6
• /
• pp.445-449
• /
• 2009

In this paper, we propose a computationally efficient multi GPU accelerated image registration technique to correct the motion difference between the pre-contrast CT image and post-contrast CTA image. Our method consists of two steps: multi GPU based image registration and a cerebrovascular visualization. At first, it computes a similarity measure considering the parallelism between both GPUs as well as the parallelism inside GPU for performing the voxel-based registration. Then, it subtracts a CT image transformed by optimal transformation matrix from CTA image, and visualizes the subtracted volume using GPU based volume rendering technique. In this paper, we compare our proposed method with existing methods using 5 pairs of pre-contrast brain CT image and post-contrast brain CTA image in order to prove the superiority of our method in regard to visual quality and computational time. Experimental results show that our method well visualizes a brain vessel, so it well diagnose a vessel disease. Our multi GPU based approach is 11.6 times faster than CPU based approach and 1.4 times faster than single GPU based approach for total processing.

• Korean Chemical Engineering Research
• /
• v.46 no.5
• /
• pp.938-944
• /
• 2008

The electro-optic properties of polymer-dispersed liquid crystal cells containing BDVE(Butanediol vinyl ether) in PN393 base pre-polymer were examined. The higher the contents of BDVE, the smaller becomes the droplet size. However, the droplet size was saturated around $3{\mu}m$ even at 40 wt% of BDVE. Both of contrast ratio and response time of PDLC cell fabricated with a new formula were found to be superior to the reference cell with PN393 by the factor of 4.9 and 0.15, respectively. However, the new formula made the operating voltage go higher compared to the reference cell of PN393 formula. Except for contrast ratio, response time as well as operating voltage were found to be highly stabilized by adding BDVE in PN393 base pre-polymer over the temperature range of $0{\sim}60^{\circ}C$ studied.

• IEIE Transactions on Smart Processing and Computing
• /
• v.3 no.2
• /
• pp.52-58
• /
• 2014

Exact histogram specification (EHS) transforms the histogram of an input image into the specified histogram. In the conventional EHS techniques, the pixels are first sorted according to their graylevels, and the pixels that have the same graylevel are further differentiated according to the local average of the pixel values and the edge strength. The strictly ordered pixels are then mapped to the desired histogram. However, since the conventional sorting method is inherently dependent on the initial graylevel-based sorting, the contrast enhancement capability of the conventional EHS algorithms is restricted. We propose a modified EHS algorithm considering the just noticeable difference. In the proposed algorithm, the edge pixels are pre-processed such that the output edge pixels obtained by the modified EHS can result in the local contrast enhancement. Moreover, we introduce a new sorting method for the pixels that have the same graylevel. Experimental results show that the proposed algorithm provides better image enhancement performance compared to the conventional EHS algorithms.

• Proceedings of the KSMRM Conference
• /
• 2002.11a
• /
• pp.135-135
• /
• 2002

Purpose： To assess diagnostic accuracy of dynamic contrast enhanced MR mammography in differentiating between benign and malignant lesions. Materials and methods： Ninety-three patients with suspicious mammographic, sonographic or palpable findings underwent pre- or postoperative contrast-enhanced MR imaging of breast using three dimensional fast low-angle shot (3D FLASH) sequence (16/4 msec［repetition time / echo time］, 20 flip angle, 3mm slice thickness with no slice gap, 256 by 256 in-plane matrix) covering whole breasts. T1 weighted images were obtained before and after bolus administration of gadopentetate dimeglumine (0.15 mmol/kg). Subtraction images and time-signal intensity curves of region of interest were obtained sequentially and correlated with pathologic diagnoses of lesions.

• Proceedings of the Korea Society for Simulation Conference
• /
• 2003.11a
• /
• pp.201-206
• /
• 2003

In clinical practice, CT Angiography is a powerful technique for the visualziation of blood flow in arterial vessels throughout the body. However CT Angiography images of blood vessels anywhere in the body may be fuzzy if the patient moves during the exam. In this paper, we propose a novel technique for removing global motion artifacts in the 3D space. The proposed methods are based on the two key ideas as follows. First, the method involves the extraction of a set of feature points by using a 3D edge detection technique based on image gradient of the mask volume where enhanced vessels cannot be expected to appear, Second, the corresponding set of feature points in the contrast volume are determined by correlation-based registration. The proposed method has been successfully applied to pre- and post-contrast CTA brain dataset. Since the registration for motion correction estimates correlation between feature points extracted from skull area in mask and contrast volume, it offers an accelerated technique to accurately visualize blood vessels of the brain.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.41-51
• /
• 2024

In this paper, we explore the enhancement of target detection accuracy in the guided weapon using deep learning object detection on infrared (IR) images. Due to the characteristics of IR images being influenced by factors such as time and temperature, it's crucial to ensure a consistent representation of object features in various environments when training the model. A simple way to address this is by emphasizing the features of target objects and reducing noise within the infrared images through appropriate pre-processing techniques. However, in previous studies, there has not been sufficient discussion on pre-processing methods in learning deep learning models based on infrared images. In this paper, we aim to investigate the impact of image pre-processing techniques on infrared image-based training for object detection. To achieve this, we analyze the pre-processing results on infrared images that utilized global or local information from the video and the image. In addition, in order to confirm the impact of images converted by each pre-processing technique on object detector training, we learn the YOLOX target detector for images processed by various pre-processing methods and analyze them. In particular, the results of the experiments using the CLAHE (Contrast Limited Adaptive Histogram Equalization) shows the highest detection accuracy with a mean average precision (mAP) of 81.9%.

• The Mathematical Education
• /
• v.55 no.4
• /
• pp.397-416
• /
• 2016

This study analyzed the quality of mathematics classes with observations using the instrument, MQI(Mathematical Quality of Instruction). Class recordings and interviews were conducted on 2 pre-service teachers and 4 in-service teachers. This study recorded and analyzed 3 or 4 classes for each mathematics teacher by using revised MQI. There were a total of 8 raters: 2 or 3 raters analyzed each class. MQI has four dimensions: Richness of the Mathematics, Working with Students and Mathematic, Errors and Imprecision, Student Participation in Meaning-Making and Reasoning. In the dimension of 'Richness of Mathematics', all teachers had good scores of 'explanations of teacher' but had lower scores of 'linking and connections', 'multiple procedures or solution methods' and 'developing mathematical generalizations.' In the dimension of 'Working with Students and Mathematics', two in-service teachers who have worked and having more experience had higher scores than others. In the dimension of 'Errors and Imprecision', all teachers had high scores. In the dimension of 'Student Participation in Meaning-Making and Reasoning', two pre-service teachers had contrast and also two in-service teachers who hadn't worked not long had contrast. Implications were deducted from finding to improving quality of mathematics classes.

• Korean Journal of Radiology
• /
• v.21 no.11
• /
• pp.1230-1238
• /
• 2020

Objective: We aimed to assess the effects of remote ischemic pre-conditioning (RIPC) on the incidence of contrast-induced nephropathy (CIN) after an intravenous (IV) or intra-arterial injection of contrast medium (CM) in patient and control groups. Materials and Methods: This prospective, randomized, single-blinded, controlled trial included 26 patients who were hospitalized for the evaluation of the feasibility of transcatheter aortic valve implantation and underwent investigations including contrast-enhanced computed tomography (CT), with Mehran risk scores greater than or equal to six. All the patients underwent four cycles of five minute-blood pressure cuff inflation followed by five minutes of total deflation. In the RIPC group (n = 13), the cuff was inflated to 50 mm Hg above the patient's systolic blood pressure (SBP); in the control group (n = 13), it was inflated to 10 mm Hg below the patient's SBP. The primary endpoint was the occurrence of CIN. Additionally, variation in the serum levels of cystatin C was assessed. Results: One case of CIN was observed in the control group, whereas no cases were detected in the RIPC group (p = 0.48, analysis of 25 patients). Mean creatinine values at the baseline, 24 hours after injection of CM, and 48 hours after injection of CM were 88 ± 32 μmol/L, 91 ± 28 μmol/L and 82 ± 29 μmol/L, respectively (p = 0.73) in the RIPC group, whereas in the control group, they were 100 ± 36 μmol/L, 110 ± 36 μmol/L, and 105 ± 34 μmol/L, respectively (p = 0.78). Cystatin C values (median [Q1, Q3]) at the baseline, 24 hours after injection of CM, and 48 hours after injection of CM were 1.10 [1.08, 1.18] mg/L, 1.17 [0.97, 1.35] mg/L, and 1.12 [0.99, 1.24] mg/L, respectively (p = 0.88) in the RIPC group, whereas they were 1.11 [0.97, 1.28] mg/L, 1.13 [1.08, 1.25] mg/L, and 1.16 [1.03, 1.31] mg/L, respectively (p = 0.93), in the control group. Conclusion: The risk of CIN after an IV injection of CM is very low in patients with Mehran risk score greater than or equal to six and even in the patients who are unable to receive preventive hyperhydration. Hence, the Mehran risk score may not be an appropriate method for the estimation of the risk of CIN after IV CM injection.

• Journal of Veterinary Clinics
• /
• v.37 no.6
• /
• pp.317-323
• /
• 2020

The aims of this study were to describe the appearance and size of the normal canine prostate using magnetic resonance (MR) imaging and to calculate the apparent diffusion coefficient (ADC) values. MR images were obtained from seven intact male beagle dogs using a 1.5 T MR unit. The sequences included pre- and post-contrast T1- and T2-weighted imaging with and without fat saturation. The signal intensity of the prostate was compared with the adjacent musculature, fat, and urine in the urinary bladder. We recorded the mean prostatic length, width, and height and the length of the sixth lumbar vertebral body (L6). In addition, the prostatic length (rL), width (rW), and height (rH) ratios to L6 were calculated. Diffusion-weighted images of the prostate were obtained and ADC values were calculated. The prostate was bilobed and oval-shaped, homogenous on T1-weighted images, and heterogeneous with radiating lines on T2-weighted images. Post-contrast T1-weighted sequences showed contrast enhancement of the central and radiating striations. The prostatic capsule was clearly identified on post-contrast T1-weighted images with fat saturation. The ADC values were 1.72-2.04 × 10-3mm2/sec (mean, 1.88 × 10-3mm2/sec). Knowledge of the normal appearance of the prostate on MR images is essential to assess prostatic diseases in dogs.

• Journal of Multimedia Information System
• /
• v.8 no.2
• /
• pp.79-84
• /
• 2021

Detecting brain tumors of different sizes is a challenging task. This study aimed to identify brain tumors using detection algorithms. Most studies in this area use segmentation; however, we utilized detection owing to its advantages. Data were obtained from 64 patients and 11,200 MR images. The deep learning model used was RetinaNet, which is based on ResNet152. The model learned three different types of pre-processing images: normal, general histogram equalization, and contrast-limited adaptive histogram equalization (CLAHE). The three types of images were compared to determine the pre-processing technique that exhibits the best performance in the deep learning algorithms. During pre-processing, we converted the MR images from DICOM to JPG format. Additionally, we regulated the window level and width. The model compared the pre-processed images to determine which images showed adequate performance; CLAHE showed the best performance, with a sensitivity of 81.79%. The RetinaNet model for detecting brain tumors through deep learning algorithms demonstrated satisfactory performance in finding lesions. In future, we plan to develop a new model for improving the detection performance using well-processed data. This study lays the groundwork for future detection technologies that can help doctors find lesions more easily in clinical tasks.