• 반도체디스플레이기술학회지
• /
• 제21권3호
• /
• pp.87-91
• /
• 2022

Bokeh effect is a stylistic technique that can produce blurring the background of photos. This paper implements to produce a bokeh effect with a single image by post processing. Generating depth map is a key process of bokeh effect, and depth map is an image that contains information relating to the distance of the surfaces of scene objects from a viewpoint. First, this work presents algorithms to determine the depth map from a single input image. Then, we obtain a sparse defocus map with gradient ratio from input image and blurred image. Defocus map is obtained by propagating threshold values from edges using matting Laplacian. Finally, we obtain the blurred image on foreground and background segmentation with bokeh effect achieved. With the experimental results, an efficient image processing method with bokeh effect applied using a single image is presented.

• 방송공학회논문지
• /
• 제27권3호
• /
• pp.437-450
• /
• 2022

생성적 적대 신경망(GAN, Generative Adversarial Networks)는 이미지 생성 분야에서 주목할 만한 발전을 이루었다. 하지만 큰 데이터 셋에서 불안정한 모습을 보인다는 한계 때문에 다양한 응용 분야에 쉽게 적용하기 어렵다. 단일 이미지 생성적 적대 신경망은 한장의 이미지의 내부 분포를 잘 학습하여 다양한 영상을 생성하는 분야이다. 큰 데이터셋이 아닌 단 한장만 학습함으로써 안정적인 학습이 가능하며 이미지 리타겟팅, 이미지 조작, super resolution 등 다양한 분야에 활용 가능하다. 본 논문에서는 SinGAN, ConSinGAN, InGAN, DeepSIM, 그리고 One-Shot GAN 총 다섯 개의 단일 이미지 생성적 적대 신경망을 살펴본다. 우리는 각각의 단일 이미지 생성적 적대 신경망 모델들의 성능을 비교하고 장단점을 분석한다.

• 한국해양공학회지
• /
• 제34권6호
• /
• pp.442-453
• /
• 2020

Underwater color images suffer from low visibility and color cast effects caused by light attenuation by water and floating particles. This study applied single image enhancement techniques to enhance the quality of underwater images and compared their performance with real underwater images taken in Korean waters. Dark channel prior (DCP), gradient transform, image fusion, and generative adversarial networks (GAN), such as cycleGAN and underwater GAN (UGAN), were considered for single image enhancement. Their performance was evaluated in terms of underwater image quality measure, underwater color image quality evaluation, gray-world assumption, and blur metric. The DCP saturated the underwater images to a specific greenish or bluish color tone and reduced the brightness of the background signal. The gradient transform method with two transmission maps were sensitive to the light source and highlighted the region exposed to light. Although image fusion enabled reasonable color correction, the object details were lost due to the last fusion step. CycleGAN corrected overall color tone relatively well but generated artifacts in the background. UGAN showed good visual quality and obtained the highest scores against all figures of merit (FOMs) by compensating for the colors and visibility compared to the other single enhancement methods.

• Journal of information and communication convergence engineering
• /
• 제6권2호
• /
• pp.204-206
• /
• 2008

In this paper, we proposed the recognition method that automatically extracts cracks from a surface image acquired by a digital camera and recognizes the directions (horizontal, vertical, -45 degree, and 45 degree) of cracks using the fuzzy single layer perceptron. We compensate an effect of light on a concrete surface image by applying the closing operation, which is one of the morphological techniques, extract the edges of cracks by Sobel masking, and binarize the image by applying the iterated binarization technique. Two times of noise reduction are applied to the binary image for effective noise elimination. After the specific regions of cracks are automatically extracted from the preprocessed image by applying Glassfire labeling algorithm to the extracted crack image, the cracks of the specific region are enlarged or reduced to $30{\times}30$ pixels and then used as input patterns to the fuzzy single layer perceptron. The experiments using concrete crack images showed that the cracks in the concrete crack images were effectively extracted and the fuzzy single layer perceptron was effective in the recognition of the extracted cracks directions.

• 한국멀티미디어학회논문지
• /
• 제24권8호
• /
• pp.1059-1067
• /
• 2021

In this paper, a HDR algorithm for a single image is proposed using the exposure fusion, that adaptively calculates gamma correction coefficients according to the image distribution. Since typical HDR methods should use at least three images with different exposure values at the same scene, the main problem was that they could not be applied at the single shot image. Thus, HDR enhancements based on a single image using tone mapping and histogram modifications were recently presented, but these created some location-specific noises due to improper corrections. Therefore, the proposed algorithm calculates proper gamma coefficients according to the distribution of the input image and generates different exposure images which are corrected by the dark and the bright region stretching. A HDR image reproduction controlling exposure fusion weights among the gamma corrected and the original pixels is presented. As the result, the proposed algorithm can reduce certain noises at both the flat and the edge areas and obtain subjectively superior image quality to that of conventional methods.

• 대한전자공학회논문지SP
• /
• 제44권5호
• /
• pp.103-111
• /
• 2007

본 논문에서는 single-chip CMOS Image Sensor(CIS)용 고화질 image signal processor(ISP)에 최적화된 하드웨어 구조를 제안한다. Single-chip CIS는 CIS와 ISP가 하나의 칩으로 구현된 것으로, 다양한 휴대기기에 사용된다. 휴대기기의 특성상, single-chip CIS용 ISP는 고화질이면서도 저전력을 위해 하드웨어 복잡도를 최소화해야 한다. 영상의 품질 향상을 위해서 다양한 영상 처리 블록들이 ISP에 적용되지만, 그 중에 핵심이면서 하드웨어 복잡도가 가장 큰 블록은 컬러 영상을 만들기 위한 색 보간 블록과 영상을 선명하게 하기 위한 화질 개선 필터 블록이다. 이들 블록은 데이터 처리를 위한 로직 외에도 라인 메모리를 필요로 하기 때문에 ISP의 하드웨어 복잡도의 대부분을 차지한다. 기존 ISP에서는 색 보간과 화질 개선 필터를 독립적으로 수행하였기 때문에 많은 수의 라인 메모리가 필요하였다. 따라서 하드웨어 복잡도를 낮추기 위해서는 낮은 성능의 색보간 알고리즘을 적용하거나, 화질 개선 필터를 사용하지 않아야 했다. 본 논문에서는 화질 개선을 위해 경계 적응적이면서 채널간 상관관계를 고려하는 고화질 색 보간 알고리즘을 적용하였다. 또한 채널 간 상관관계를 고려하는 색 보간 알고리즘의 특성을 이용하여 색 보간 블록과 화질 개선 필터 블록이 라인 메모리를 공유하도록 설계함으로써, 전체 라인 메모리 수를 최소화하는 새로운 구조를 제안한다. 제안된 방법을 적용하면 화질 개선 필터 블록을 위한 추가적인 라인 메모리가 불필요하기 때문에, 고화질과 낮은 복잡도 모두를 만족시킬 수 있다. 제안 방식과 기존 방식의 MSE(Mean Square Error)는 0.37로, 메모리 공유로 인한 화질의 저하는 거의 없었고, 고화질 색 보간 알고리즘을 적용했기 때문에 전체적인 화질은 향상되었다. 제안된 ISP 구조는 Verilog HDL 및 FPGA를 이용하여 실시간으로 구현 검증되었다. 0.25um CMOS 표준 셀 라이브러리를 이용하여 합성하였을 때, 총 게이트 수는 37K개였으며 7.5개의 라인 메모리가 사용되었다.

• International Journal of Computer Science & Network Security
• /
• 제21권11호
• /
• pp.17-22
• /
• 2021

Recently, image super-resolution techniques used in convolutional neural networks (CNN) have led to remarkable performance in the research area of digital image processing applications and computer vision tasks. Convolutional layers stacked on top of each other can design a more complex network architecture, but they also use more memory in terms of the number of parameters and introduce the vanishing gradient problem during training. Furthermore, earlier approaches of single image super-resolution used interpolation technique as a pre-processing stage to upscale the low-resolution image into HR image. The design of these approaches is simple, but not effective and insert the newer unwanted pixels (noises) in the reconstructed HR image. In this paper, authors are propose a novel single image super-resolution architecture based on synchronized depthwise separable convolution with Dense Skip Connection Block (DSCB). In addition, unlike existing SR methods that only rely on single path, but our proposed method used the synchronizes path for generating the SISR image. Extensive quantitative and qualitative experiments show that our method (SDCN) achieves promising improvements than other state-of-the-art methods.

• 대한임베디드공학회논문지
• /
• 제14권4호
• /
• pp.197-205
• /
• 2019

In this paper, an efficient single dehazing algorithm is proposed based on linear transformation by assuming that a linear relationship exists in saturation component between the haze image and haze-free image. First, we analyze the linearity of saturation channel, estimate the medium transmission map in terms of the saturation component. Then, the intensity of haze-free image is assumed by using CLAHE to enhance contrast of haze image. Experimental results demonstrate that proposed algorithm can naturally recover the image, especially can remove color distortion caused by conventional methods. Therefore, our approach is competitive with other state-of-the art single dehazing methods.

• 반도체디스플레이기술학회지
• /
• 제18권4호
• /
• pp.154-159
• /
• 2019

Undesirable negative image is occurred in photographs taken across partial reflections such as glass window and electronic display. Efficient removing reflections given a single image are in the spotlight in recent researches. This paper discusses and evaluates two published image reflection removal algorithms, and compares the performance of time and quality of those methods with a common dataset. As benchmarking test cases are presented, we propose to modify one of the methods to reduce the run-time with small effects on the similar image quality.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1996년도 춘계학술대회 논문집
• /
• pp.135-139
• /
• 1996

In this a paper, a new method to measure the wear of the single crystal diamond(SCD) tool using image processing is presented. To increase resoultion, high magnifying lens is used and to enlarge the measurement field of view, a image region matching method is applied. The shape of SCD tool is modeled by mathematical analysis. Cutting edge chipping and wear are calculated by the model. This method is proved to be efficient in detecting a few micron of wear and cutting edge loss by chipping along the whole cutting edge.