• Journal of Internet Computing and Services
• /
• v.19 no.3
• /
• pp.35-42
• /
• 2018

Deep learning is known as a method of high accuracy among several methods for image classification. In this paper, we propose a method of enhancing the accuracy of image classification using CNN with a batch normalization method for classification of images using deep CNN (Convolutional Neural Network). In this paper, we propose a method to add a batch normalization layer to existing neural networks to enhance the accuracy of image classification. Batch normalization is a method to calculate and move the average and variance of each batch for reducing the deflection in each layer. In order to prove the superiority of the proposed method, Accuracy and mAP are measured by image classification experiments using five image data sets SHREC13, MNIST, SVHN, CIFAR-10, and CIFAR-100. Experimental results showed that the CNN with batch normalization is better classification accuracy and mAP rather than using the conventional CNN.

• Journal of The Korean Association of Information Education
• /
• v.19 no.4
• /
• pp.441-448
• /
• 2015

This paper describes about a software implementation for class analysis and quantization based on our video synchronization method. We proposed a new indexing method, synchronization strategies, and data structure for our analyzer implementation. We implemented a class video analyzer using intelligent multimedia technologies which can play class video selectively. Our proposed method analyzes class videos depending on the time schedule composed of introduction, development and summary stages. We apply our analysis filters to the class videos in the predefined regular intervals. We experimented on the synchronization performance of our proposed method and software. In the experimental, we could demonstrate the effectiveness and practicality of our class analyzing method within the margin of error.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2009.10a
• /
• pp.211-212
• /
• 2009

• Journal of Broadcast Engineering
• /
• v.14 no.4
• /
• pp.450-462
• /
• 2009

So far, many methods for segmenting single images or video have been proposed, but few methods have dealt with multiple images with analogous content. These images, which we term consistent scene images, include concurrent images of a scene and gathered images of a similar foreground, and may be collectively utilized to describe a scene or as input images for multi-view stereo. In this paper, we present a method to segment these images with minimum user input, specifically, manual segmentation of one image, by iteratively propagating information via multi-level cues with adaptive confidence depending on the nature of the images. Propagated cues are used as the bases to compute multi-level potentials in an MRF framework, and segmentation is done by energy minimization. Both cues and potentials are classified as low-, mid-, and high- levels based on whether they pertain to pixels, patches, and shapes. A major aspect of our approach is utilizing mid-level cues to compute low- and mid- level potentials, and high-level cues to compute low-, mid-, and high- level potentials, thereby making use of inherent information. Through this process, the proposed method attempts to maximize the amount of both extracted and utilized information in order to maximize the consistency of the segmentation. We demonstrate the effectiveness of the proposed method on several sets of consistent scene images and provide a comparison with results based only on mid-level cues [1].

• Proceedings of the KOSOMBE Conference
• /
• v.1985 no.06
• /
• pp.22-25
• /
• 1985

본고에서, "나선형 주사 방법에 의한 고속 NMR 영상화" 방법을 제안하고 그에 따른 실험 결과를 보였다. 이것은 2차원 FID 영역을 나선형 궤적으로 스캐닝하며 데이타를 받을 수 있도록 경사 자계 파형 (gradient pulse)을 가하여 빠른 시간에 (수십 msec - 수초)내에 영상 정보를 얻어낸 후, 재구성 알고리즘을 씀으로써 영상을 얻어내는 방법이다. 이 방법의 장점은 첫째로 $T_2$ 감쇄에 의한 PSF (Point Spread Function)가 윈형 대칭으로 주어지므로 영상화 했을때 물체의 구조 식별이 기존의 EPI (Echo Planar Imaging) 방법에 비해서 선명하며, 둘째로 나선형 궤적을 구현하기 의한 경사 자계 파형에서 불연속 점을 없앰과 동시에 파형의 세기가 점차로 증가하는 형태이므로, 기존의, 파형 왜곡에 의해 영상에 미치는 영향을 최소화 할 수가 있으며, 세째로 나선형 스캔을 사이 사이에 끼워 넣는 방법을 씀으로써 해상도를 향상시킬 수가 있다.

• The KIPS Transactions:PartB
• /
• v.15B no.5
• /
• pp.383-390
• /
• 2008

Block based transform coding is most popular approach for image and video compression. However it suffers from severe quality degradation especially from blocking artifacts. The subjective quality degradation caused by such blocking artifacts in general does not agree well with an objecive quality measurement such as PSNR. Hence new quality evaluation technique is necessary. We propose a new image quality assessment method by measuring blocking artifacts for block based transform coded images. In order to characterize blocking artifacts, proposed method utilizes the facts that, blocking artifacts, when occur, have different pixel values along the block boundaries and such differences usually continuously span along the whole boundaries. This method does not require the original uncompressed image. It operates on single block boundary and quantifies the amount of blocking artifacts on it. Experiments on various compressed images various bitrates show that proposed quantitative measure of blocking artifacts matches well with the subjective quality of them judged by human visual system.

• The Journal of the Korea Contents Association
• /
• v.15 no.11
• /
• pp.18-26
• /
• 2015

Geometric distortion and shadow effect due to capturing angle could be included in document copy images that are captured by a camera in stead of a scanner. In this paper, a clean black and white document image generation algorithm by distortion correction and shadow elimination based on a camera, is proposed. In order to correct geometric distortion such as straightening un-straight boundary lines occurred by camera lens radial distortion and eliminating outlying area included by camera direction, second derivative filter based document boundary detection method is developed. Black and white images have been generated by adaptive binarization method by eliminating shadow effect. Experimental results of the black and white document image generation algorithm by recovering geometrical distortion and eliminating shadow effect for the document images captured by smart phone camera, shows very good processing results.

• Journal of Korea Multimedia Society
• /
• v.9 no.7
• /
• pp.808-817
• /
• 2006

Although the generally used the velocity index of doppler effect is a very significant factor in the functional evaluation of the left ventricle, it depends on the subjective evaluation of an inspector. The objective data of the motility can be obtained from the gated myocardial SPECT images by quantitative analysis. However, it is difficult to image visual of the velocity of the motion. The aim of our study is to develop a new method for the imaging velocity using the gated myocardial SPECT images and use it as an evaluation index for analyzing motility. First we visualized left ventricle into 3 dimensions using the coordinates of the points which were obtained through a segmentation of myocardium. Each point was represented by the different colors, according to the velocity of each point. We performed a validation study using 7 normal subjects and 15 myocardial infarction patients. To analyze motility, we used the average of the moved distance and the velocity. In normal cases, the average of the moved distance was 4.3mm and the average of the velocity was 11.9mm. In patient cases, the average of the moved distance was 3.9mm and the average of the velocity was 10.5mm. These results show that the motility of normal subjects is higher than the abnormal subjects. We expect that our proposed method could become a way to improve the accuracy and reproducibility for the functional evaluation of myocardial wall.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.39 no.5
• /
• pp.11-20
• /
• 2002

Based on regularization principles, we propose a new dequantization scheme on DCT-based transform coding for reducing of blocking artifacts and minimizing the quantization error. The conventional image dequantization is simply to multiply the received quantized DCT coefficients by the quantization matrix. Therefore, for each DCT coefficients, we premise that the quantization noise is as large as half quantizer step size (in DCT domain). Our approach is based on basic constraint that quantization error is bounded to ${\pm}$(quantizer spacing/2) and at least there are not high frequency components corresponding to discontinuities across block boundaries of the images. Through regularization, our proposed dequantization scheme, sharply reduces blocking artifacts in decoded images. Our proposed algorithm guarantees that the dequantization process will map the quantized DCT coefficients will be evaluated against the standard JPEG, MPEG-1 and H.263 (with Annex J deblocking filter) decoding process. The experimental results will show visual improvements as well as numerical improvements in terms of the peak-signal-to-noise ratio (PSNR) and the blockiness measure (BM) to be defined.

• Journal of the Korean Society of Radiology
• /
• v.84 no.5
• /
• pp.1091-1093
• /
• 2023