• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2002년도 가을 학술발표논문집 Vol.29 No.2 (2)
• /
• pp.301-303
• /
• 2002

Subspace analysis is a popular method for multivariate data analysis and is closely related to factor analysis and principal component analysis (PCA). In the context of image processing (especially positron emission tomography), all data points are nonnegative and it is expected that both basis images and factors are nonnegative in order to obtain reasonable result. In this paper We present a sequential EM algorithm for rectified subspace analysis (subspace in nonnegativity constraint) and apply it to dynamic PET image analysis. Experimental results show that our proposed method is useful in dynamic PET image analysis.

• International Journal of Advanced Culture Technology
• /
• 제10권2호
• /
• pp.252-259
• /
• 2022

Artificial intelligence is used in fusion with image processing techniques using cameras. Image processing technology is a technology that processes objects in an image received from a camera in real time, and is used in various fields such as security monitoring and medical image analysis. If such image processing reduces the accuracy of recognition, providing incorrect information to medical image analysis, security monitoring, etc. may cause serious problems. Therefore, this paper uses a mixture of YOLOv4-tiny model and image processing algorithm and uses the COCO dataset for learning. The image processing algorithm performs five image processing methods such as normalization, Gaussian distribution, Otsu algorithm, equalization, and gradient operation. For RGB images, three image processing methods are performed: equalization, Gaussian blur, and gamma correction proceed. Among the nine algorithms applied in this paper, the Equalization and Gaussian Blur model showed the highest object detection accuracy of 96%, and the gamma correction (RGB environment) model showed the highest object detection rate of 89% outdoors (daytime). The image binarization model showed the highest object detection rate at 89% outdoors (night).

• 한국컴퓨터정보학회지
• /
• 제12권1호
• /
• pp.9-19
• /
• 2004

본 논문에서는 내용 기반 영상 검색 시스템(Content Based Image Retrieval System)의 특징 추출(feature extraction)과 분석(analysis)을 위한 방법으로 적응적 컴포넌트 분석(ACA: Adaptive Component Analysis)을 제안하고 있다. 검색을 위해서 영상에서 추출된 특징들은 영상의 도메인(domain)에 따라 적절하게 적용해야만 좋은 검색 결과를 얻을 수 있다. 이러한 조건을 만족시키기 위한 방법으로 본 논문에서는 검색 측정도(retrieval measurement)를 제안하고 있다. ACA는 알고리즘과 시스템적인 관점에서 볼 때, 기존의 내용 기반 영상 검색을 위한 중간 단계라고 할 수 있으며, 검색 속도 향상 및 성능 개선에 목표를 두고 있다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제15권1호
• /
• pp.35-44
• /
• 2015

The main purpose of image enhancement is to improve certain characteristics of an image to improve its visual quality. This paper proposes a method for image contrast enhancement that can be applied to both medical and natural images. The proposed algorithm is designed to achieve contrast enhancement while also preserving the local image details. To achieve this, the proposed method combines local image contrast preserving dynamic range compression and contrast limited adaptive histogram equalization (CLAHE). Global gain parameters for contrast enhancement are inadequate for preserving local image details. Therefore, in the proposed method, in order to preserve local image details, local contrast enhancement at any pixel position is performed based on the corresponding local gain parameter, which is calculated according to the current pixel neighborhood edge density. Different image quality measures are used for evaluating the performance of the proposed method. Experimental results show that the proposed method provides more information about the image details, which can help facilitate further image analysis.

• 전자공학회논문지B
• /
• 제30B권9호
• /
• pp.54-63
• /
• 1993

Quantization error, generated by the quantization process of an image, is inherent in computer vision. Because, especially in stereo vision, the quantization error in a 2-D image results in position errors in the reconstructed 3-D scene, it is necessary to analyze it mathematically. In this paper, the analysis of the probability density function (pdf) of quantization error for a line-based stereo matching scheme is presented. We show that the theoretical pdf of quantization error in the reconstructed 3-D position information has more general form than the conventional analysis for pixel-based stereo matching schemes. Computer simulation is observed to surpport the theoretical distribution.

• 반도체디스플레이기술학회지
• /
• 제15권1호
• /
• pp.27-32
• /
• 2016

A new lane detection algorithm is proposed for the analysis of DNA fingerprints from a polymerase chain reaction (PCR) gel electrophoresis image. Although several research results have been previously reported, it is still challenging to extract lanes precisely from images having abrupt background brightness difference and bent lanes. We propose an edge based algorithm for calculating the average lane width and lane cycle. Our method adopts sub-pixel algorithm for extracting rising-edges and falling edges precisely and estimates the lane width and cycle by using k-means clustering algorithm. To handle the curved lanes, we partition the gel image into small portions, and track the lane centers in each partitioned image. 32 gel images including 534 lanes are used to evaluate the performance of our method. Experimental results show that our method is robust to images having background difference and bent lanes without any preprocessing.

• 치과방사선
• /
• 제27권1호
• /
• pp.7-16
• /
• 1997

This study was performed to demonstrate the method of image reformation for dental implants, using a personal computer with inexpensive softwares and to compare the images reformatted using the above method with those using Dentascan software. CT axial slices of 5 mandibles of 5 volunteers from GE Highspeed Advantage(GE Medical systems, U.S.A.) were used. Personal computer used for image reformation was PowerWave 6041120 (Power Computing Co, U.S.A.) and softwares used were Osiris (Univ. Hospital of Geneva, Switzerland) and Import ACCESS V1.H Designed Access Co., U.S.A.) for importing CT images and NIH Image 1.58 (NIH, U.S.A.) for image processing. Seven images were selected among the serial reconstructed cross-sectional images produced by Dentascan(DS group). Seven resliced cross-sectional images at the same position were obtained at the personal computer(PC group). Regression analysis of the measurements of PC group was done against those of DS group. Measurements of the bone height and width at the reformed cross-sectional images using Mac-compatible computer were highly correlated with those using workstation with Dentascan software(height : r²=0.999, p<0.001, width : r²=0.991, p<0.001). So, it is considered that we can use a personal computer with inexpensive softwares for the dental implant planning, instead of the expensive software and workstation.

• Journal of Information Processing Systems
• /
• 제11권4호
• /
• pp.601-615
• /
• 2015

Region of interest (ROI) is the most informative part of a medical image and mostly has been used as a major part of watermark. Various shapes ROIs selection have been reported in region-based watermarking techniques. In region-based watermarking schemes an image region of non-interest (RONI) is the second important part of the image and is used mostly for watermark encapsulation. In online healthcare systems the ROI wrong selection by missing some important portions of the image to be part of ROI can create problem at the destination. This paper discusses the complete medical image availability in original at destination using the whole image as a watermark for authentication, tamper localization and lossless recovery (WITALLOR). The WITALLOR watermarking scheme ensures the complete image security without of ROI selection at the source point as compared to the other region-based watermarking techniques. The complete image is compressed using the Lempel-Ziv-Welch (LZW) lossless compression technique to get the watermark in reduced number of bits. Bits reduction occurs to a number that can be completely encapsulated into image. The watermark is randomly encapsulated at the least significant bits (LSBs) of the image without caring of the ROI and RONI to keep the image perceptual degradation negligible. After communication, the watermark is retrieved, decompressed and used for authentication of the whole image, tamper detection, localization and lossless recovery. WITALLOR scheme is capable of any number of tampers detection and recovery at any part of the image. The complete authentic image gives the opportunity to conduct an image based analysis of medical problem without restriction to a fixed ROI.

• 한국멀티미디어학회논문지
• /
• 제18권10호
• /
• pp.1180-1188
• /
• 2015

In this paper, a parallel cell contour line extraction algorithm using CUDA, which has no inner contour lines, is proposed. The contour of a cell is very important in a cell image analysis. It could be obtained by a conventional serial contour tracing algorithm or parallel morphology operation. However, the cell image has various damages in acquisition or dyeing process. They could be turn into several inner contours, which make a cell image analysis difficult. The proposed algorithm introduces a min-max coordinates table into each CUDA thread block, and removes the inner contour in parallel. It is 4.1 to 7.6 times faster than a conventional serial contour tracing algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권12호
• /
• pp.4502-4512
• /
• 2014

This article presents a reformulation of the Grey Edge framework for colour constancy. Colour constancy is the ability of a visual system to perceive objects' colours independently of their scenes' illuminants. Colour constancy algorithms try to estimate the colour of an illuminant from image values. This estimation can later be used to correct the image as though it were taken under a white illuminant. The modification presented allows the framework to incorporate image-specific filters instead of the commonly used edge detectors. A colour constancy algorithm is proposed using PCA and FastICA linear component analyses methods for the construction of such filters. The results show that the proposed method improves the accuracies of the Grey Edge framework algorithms whilst on the other hand, achieving comparable accuracies with the state-of-the-art methods, but improving their time efficiencies.