• Korean Journal of Radiology
• /
• 제25권4호
• /
• pp.374-383
• /
• 2024

Objective: To evaluate the diagnostic performance and image quality of 1.5-mm slice thickness MRI with deep learningbased image reconstruction (1.5-mm MRI + DLR) compared to routine 3-mm slice thickness MRI (routine MRI) and 1.5-mm slice thickness MRI without DLR (1.5-mm MRI without DLR) for evaluating temporal lobe epilepsy (TLE). Materials and Methods: This retrospective study included 117 MR image sets comprising 1.5-mm MRI + DLR, 1.5-mm MRI without DLR, and routine MRI from 117 consecutive patients (mean age, 41 years; 61 female; 34 patients with TLE and 83 without TLE). Two neuroradiologists evaluated the presence of hippocampal or temporal lobe lesions, volume loss, signal abnormalities, loss of internal structure of the hippocampus, and lesion conspicuity in the temporal lobe. Reference standards for TLE were independently constructed by neurologists using clinical and radiological findings. Subjective image quality, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were analyzed. Performance in diagnosing TLE, lesion findings, and image quality were compared among the three protocols. Results: The pooled sensitivity of 1.5-mm MRI + DLR (91.2%) for diagnosing TLE was higher than that of routine MRI (72.1%, P < 0.001). In the subgroup analysis, 1.5-mm MRI + DLR showed higher sensitivity for hippocampal lesions than routine MRI (92.7% vs. 75.0%, P = 0.001), with improved depiction of hippocampal T2 high signal intensity change (P = 0.016) and loss of internal structure (P < 0.001). However, the pooled specificity of 1.5-mm MRI + DLR (76.5%) was lower than that of routine MRI (89.2%, P = 0.004). Compared with 1.5-mm MRI without DLR, 1.5-mm MRI + DLR resulted in significantly improved pooled accuracy (91.2% vs. 73.1%, P = 0.010), image quality, SNR, and CNR (all, P < 0.001). Conclusion: The use of 1.5-mm MRI + DLR enhanced the performance of MRI in diagnosing TLE, particularly in hippocampal evaluation, because of improved depiction of hippocampal abnormalities and enhanced image quality.

• 대한디지털의료영상학회논문지
• /
• 제9권2호
• /
• pp.39-43
• /
• 2007

In angiography, the global standard agreements of DICOM is lossless. But it brings on overload and takes too much store space in DICOM sever. Because of all those things we transmit images which is classified in subjective way. But this cause data loss and would be lead doctors to make wrong reading. As a result of that we try to transmit continued image (raw data) to reduce those mistakes. We got angiography images from the equipment(Allura FD20-Philips). And compressed it in two different methods(lossless & lossy fair). and then transmitted them to PACS system. We compared the quality of QC phantom images that are compressed by different compress method and compared spatial resolution of each images after CD copy. Then compared each Image's data volume(lossless & lossy fair). We measured spatial resolution of each image. All of them had indicated 401p/mm. We measured spatial resolution of each image after CD copy. We got also same conclusion (401p/mm). The volume of continued image (raw data) was 127.8MB(360.5 sheets on average) compressed in lossless and 29.5MB(360.5 sheets) compressed in lossy fair. In case of classified image, it was 47.35MB(133.7 sheets) in lossless and 4.5MB(133.7 sheets) in lossy fair. In case of angiography the diagnosis is based on continued image(raw data). But we transmit classified image. Because transmitting continued image causes some problems in PACS system especially transmission and store field. We transmit classified image compressed in lossless But it is subjective and would be different depend on radiologist. therefore it would make doctors do wrong reading when patients transfer another hospital. So we suggest that transmit continued image(raw data) compressed in lossy fair. It reduces about 60% of data volume compared with classified image. And the image quality is same after CD copy.

• 한국디자인학회:학술대회논문집
• /
• 한국디자인학회 2001년도 디자인학연구 2001 봄 학술발표회 논문집
• /
• pp.108-109
• /
• 2001

• 대한전자공학회논문지
• /
• 제24권2호
• /
• pp.301-308
• /
• 1987

In this paper, we have defined the Weighted Hadamard Transform (WHT) and developed efficient algorithms for the fast computation of the WHT. The WHT is applied to digital image processing and compared with Hadamard Transform (HT). We have weighted at the center spatial frequency domains of the Hadamard Transform and transmitted a image and then center high frequencies are neglected at the receiving. The WHT of signal to noise ratio(SNR) and image quality are enhanced than the HT.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.940-942
• /
• 2003

In an attempt to enhance the visual effect of SPOT image, this study develops a neural network algorithm to transform SPOT false color into simulated true color. The method has been tested using Landsat TM and SPOT images. The qualitative and quantitative comparisons indicate that the striking similarity can be found between the true and simulated true images in terms of the visual looks and the statistical analysis.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.60.2-60
• /
• 2001

In this paper, manufacturing system and Internet are combined and NC milling machine engraves image and text on nameplate. Image and text are input through Internet. And NC tool path is obtained by thinning algorithm and NC part program is generated. Thinning algorithm detects center lines from image and text by using connectivity and tool path is obtained along the center line. Actually experiments are performed and thinning algorithm and G-code generation module are verified.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.458-460
• /
• 2003

This paper proposes new algorithms for cloudy area detection by GHA (Generalized Hebbian Algorithm) and SOFM (Self-Organized Feature Map). SOFM and GHA are unsupervised neural networks and are used for pattern classification and shape detection of satellite image. Proposed algorithm is based on block based image processing that size is 16${\times}$16. Results of proposed algorithm shows good performance of cloudy area detection except blur cloudy area.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2004년도 ICEIC The International Conference on Electronics Informations and Communications
• /
• pp.427-434
• /
• 2004

In this paper, we present a overview of content-based image retrieval (CBIR) systems: its results and its problems. We propose our CBIR system currently based on color and texture. From the CBIR systems. we discuss the way to add semantic values in image retrieval systems. There are 3 ways for adding them: concept definition, machine learning and man-machine interaction. Along with this we introduce our preliminary results and discuss them in the goal of reaching semantic retrieval. Different result representation schemes are presented. At last, we present our work to build a complete annotated image database and our image annotaion program.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권7호
• /
• pp.2328-2344
• /
• 2022

2D human pose estimation still faces difficulty in low-resolution images. Most existing top-down approaches scale up the target human bonding box images to the large size and insert the scaled image into the network. Due to up-sampling, artifacts occur in the low-resolution target images, and the degraded images adversely affect the accurate estimation of the joint positions. To address this issue, we propose a multi-resolution input feature fusion network for human pose estimation. Specifically, the bounding box image of the target human is rescaled to multiple input images of various sizes, and the features extracted from the multiple images are fused in the network. Moreover, we introduce a guiding channel which induces the multi-resolution input features to alternatively affect the network according to the resolution of the target image. We conduct experiments on MS COCO dataset which is a representative dataset for 2D human pose estimation, where our method achieves superior performance compared to the strong baseline HRNet and the previous state-of-the-art methods.

• 대한의용생체공학회:의공학회지
• /
• 제27권6호
• /
• pp.392-401
• /
• 2006

Image Guided Surgery (IGS) system which has variously tried in medical engineering fields is able to give a surgeon objective information of operation process like decision making and surgical planning. This information is displayed through 3D images which are acquired from image modalities like CT and MRI for pre-operation. The technique of image registration is necessary to construct IGS system. Image registration means that 3D model and the object operated by a surgeon are matched on the common frame. Major techniques of registration in IGS system have been used by recognizing fiducial markers placed on the object. However, this method has been criticized due to additional trauma, its invasive protocol inserting fiducial markers in patient's bone and generating noise data when 2D slice images are acquired by image modality because many markers are made of metal. Therefore, this paper developed shape-based registration technique to improve the limitation of fiducial marker based IGS system. Iterative Closest Points (ICP) algorithm was used to match corresponding points and quaternion based rotation and translation transformation using closed form solution applied to find the optimized cost function of transformation. we assumed that this algorithm were used in Total Knee replacement (TKR) operation. Accordingly, we have developed region-based 3D registration technique based on anatomical landmarks and this registration algorithm was evaluated in a femur model. It was found that region-based algorithm can improve the accuracy in 3D registration.