• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.359-360
• /
• 2021

영상의 픽셀기반처리는 한 픽셀의 값을 변환하는데 다른 픽셀의 값에 관계없이 단지 현재의 픽셀의 값에만 의존하여 변환하는 처리를 의미한다. 픽셀기반 처리는 영상 변환, 영상 개선, 영상 합성 등의 많은 분야에서 가장 기초적인 연산으로 사용된다. 산술연산, 히스토그램 평활화, 명암대비 스트레칭 등의 처리 방법들이 있다. 본 논문에서는 드론으로 촬영된 서해안 갯벌 영상에서 갯벌 영역을 명확하게 구분하기 위하여 전처리 과정 중 경계검출부분에서 픽셀기반처리를 이용하여 효율적인 윤곽선을 찾기 위한 방법을 모색한다.

• 한국시뮬레이션학회논문지
• /
• 제10권3호
• /
• pp.59-70
• /
• 2001

Typical low-level image processing tasks require thousands of operations per pixel for each input image. Traditional general-purpose computers are not capable of performing such tasks in real time. Yet important features of traditional computers are not exploited by low-level image processing tasks. Since storage requirements are limited to a small number of low-precision integer values per pixel, large hierarchical memory systems are not necessary. The mismatch between the demands of low-level image processing tasks and the characteristics of conventional computers motivates investigation of alternative architectures. The structure of the tasks suggests employing an array of processing elements, one per pixel, sharing instructions issued by a single controller. In this paper we implemented various image processing filtering using the format converter. Also, we realized from conventional gray image process to color image process. This design method is based on realized the large processor-per-pixel array by integrated circuit technology This format converter design has control path implementation efficiently, and can be utilize the high technology without complicated controller hardware.

• 한국축산시설환경학회지
• /
• 제7권3호
• /
• pp.183-190
• /
• 2001

본 연구는 영상처리를 통한 보다 간편하고 정확한 젖소의 체중측정을 위해 수행되었다. 카메라와 개인용 컴퓨터를 이용하여 젖소의 영상을 받은 후 이의 화소 값들을 이용하였으며, 또한 여러 가지 방법으로 체중과의 관계를 회귀분석 방법을 하였다. 1. 본 실험의 결과 다중회귀식에 의한 계산체중, 화소표면적, 부피와 체증과의 상관계수는 각각 0.9424, 0.9439, 0.9651로 나타났으며, 젖소의 체중과 상관관계도가 높았다. 2. 실험에 이용된 젖소의 개체수를 50두이었지만, 각각 상태별 체형의 차이고 인하여 일관된 회귀식을 적용하기에는 정확도 문제에서 다소 문제가 있었다. 그래서 각 개체군으로 나누어서 연구할 필요가 있을 것으로 판단되었다. 3. 젖소의 체중 계측시간은 10초에 불과하지만 실험장치까지의 유도과정은 편균 한 마리에 10분 정도가 소요되었다.

• 대한임베디드공학회논문지
• /
• 제11권5호
• /
• pp.277-285
• /
• 2016

Moving object detection system has been an emerging research field in various advanced driver assistance systems (ADAS) and surveillance system. In this paper, we propose two optical flow based moving object detection methods at dynamic scenes. Both proposed methods consist of three successive steps; pre-processing, foreground segmentation, and post-processing steps. Two proposed methods have the same pre-processing and post-processing steps, but different foreground segmentation step. Pre-processing calculates mainly optical flow map of which each pixel has the amplitude of motion vector. Dense optical flows are estimated by using Farneback technique, and the amplitude of the motion normalized into the range from 0 to 255 is assigned to each pixel of optical flow map. In the foreground segmentation step, moving object and background are classified by using the optical flow map. Here, we proposed two algorithms. One is Gaussian mixture model (GMM) based background subtraction, which is applied on optical map. Another is adaptive thresholding based foreground segmentation, which classifies each pixel into object and background by updating threshold value column by column. Through the simulations, we show that both optical flow based methods can achieve good enough object detection performances in dynamic scenes.

• 한국정보처리학회논문지
• /
• 제6권11호
• /
• pp.3140-3149
• /
• 1999

Skew correction for document images can be using a rotational transformation of pixel coordinates. In this paper we propose a method which corrects the document skew, by an amount of $\theta$ degrees, using block information, where the block is defined as a rectangular area containing adjacent black pixels. Processing speed of the proposed method is faster than that of the method using pixel transformation, since the number of floating-point operations can be reduced significantly. In the proposed method, we rotate only the four corner points of each block, and then identify the pixels inside the block. Two methods for inside pixel identification are proposed; the first method finds two points intersecting the boundary of the rotated block in each row, and determines the pixels between the two intersection points as the inside pixel. The second method finds boundary points based on Bresenham's line drawing algorithm, using fixed-point operation, and fills the region surrounded by these boundaries as black pixels. We have measured the performance of the proposed method by experimenting it with 2,016 images of various English and Korean documents. We have also proven the superiority of our algorithm through performance comparison with respect to existing methods based on pixel transformation.

• Journal of Advanced Marine Engineering and Technology
• /
• 제28권4호
• /
• pp.587-592
• /
• 2004

Pixel classification is one of basic issues of image processing. The general characteristics of the pixels belonging to various classes are discussed and the radical principles of pixel classification are given. At the same time, a pixel classification scheme based on image information scales is proposed. The proposed method is overcome that computation amount become greater and contents easily get turned. And image directional scales has excellent anti-noise performance. In the result of experiment. good efficiency is showed compare with other methods.

• Current Optics and Photonics
• /
• 제6권5호
• /
• pp.463-472
• /
• 2022

At certain wavelengths, single-pixel imaging is considered to be a solution that can achieve high quality imaging and also reduce costs. However, achieving imaging of complex scenes is an overhead-intensive process for single-pixel imaging systems, so low efficiency and high consumption are the biggest obstacles to their practical application. Improving efficiency to reduce overhead is the solution to this problem. Salient object detection is usually used as a pre-processing step in computer vision tasks, mimicking human functions in complex natural scenes, to reduce overhead and improve efficiency by focusing on regions with a large amount of information. Therefore, in this paper, we explore the implementation of salient object detection based on single-pixel imaging after a single pixel, and propose a scheme to reconstruct images based on Fourier bases and use U²Net models for salient object detection.

• Journal of Information Processing Systems
• /
• 제16권1호
• /
• pp.49-60
• /
• 2020

To solve the edge ringing or block effect caused by the partial differential diffusion in image enhancement domain, a new image enhancement algorithm based on bidirectional diffusion, which smooths the flat region or isolated noise region and sharpens the edge region in different types of defect images on aviation composites, is presented. Taking the image pixel's neighborhood intensity and spatial characteristics as the attribute descriptor, the presented bidirectional diffusion model adaptively chooses different diffusion criteria in different defect image regions, which are elaborated are as follows. The forward diffusion is adopted to denoise along the pixel's gradient direction and edge direction in the pixel's smoothing area while the backward diffusion is used to sharpen along the pixel's gradient direction and the forward diffusion is used to smooth along the pixel's edge direction in the pixel's edge region. The comparison experiments were implemented in the delamination, inclusion, channel, shrinkage, blowhole and crack defect images, and the comparison results indicate that our algorithm not only preserves the image feature better but also improves the image contrast more obviously.

• 전기전자학회논문지
• /
• 제18권3호
• /
• pp.403-408
• /
• 2014

본 논문은 Varying Interpolator를 개선하여 다수의 Pixel을 한 번에 처리할 수 있는 Rasterizer 구조를 제안한다. 설계한 Rasterizer의 Varying Interpolator는 한 번에 16 Pixel을 처리 할 수 있으며 최대 64개의 색상을 출력으로 가진다. 또한 Rasterizer의 연산을 행렬연산 및 행렬변환으로 구성하여 연산의 중복성을 줄이고 재사용성을 높여 Rasterizer의 처리 속도를 높였다. 제안하는 구조의 Rasterizer 는 기존의 연구와 비교하여 색상 보간은 11%, Rasterizer 전체 처리 속도는 17% 향상된 성능을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권5호
• /
• pp.1671-1686
• /
• 2022

Face manipulation tools represented by Deepfake have threatened the security of people's biological identity information. Particularly, manipulation tools with deep learning technology have brought great challenges to Deepfake detection. There are many solutions for Deepfake detection based on traditional machine learning and advanced deep learning. However, those solutions of detectors almost have problems of poor performance when evaluated on different quality datasets. In this paper, for the sake of making high-quality Deepfake datasets, we provide a preprocessing method based on the image pixel matrix feature to eliminate similar images and the residual channel attention network (RCAN) to resize the scale of images. Significantly, we also describe a Deepfake detector named Cascaded-Hop which is based on the PixelHop++ system and the successive subspace learning (SSL) model. By feeding the preprocessed datasets, Cascaded-Hop achieves a good classification result on different manipulation types and multiple quality datasets. According to the experiment on FaceForensics++ and Celeb-DF, the AUC (area under curve) results of our proposed methods are comparable to the state-of-the-art models.