Inpainting GAN-Based Image Blending with Adaptive Binary Line Mask

Thanh Hien Truong,이태호,Viduranga Munasinghe,김태성,김진성,이혁재 한국멀티미디어학회 2023 The journal of multimedia information system Vol.10 No.3

Image blending is a scheme for image composition to make the composite image looks as natural and realistic as possible. Image blending should ensure that the edges of the object look seamless and do not distort colors. Recently, numerous studies investigated image blending methods adopting deep learning-based image processing algorithms and contributed to generating natural blended images. Although the previ-ous studies show remarkable performance in many cases, they suffer from quality drop when blending incompletely cropped object. This is because partial loss and unnecessary extra information on the cropped object image interferes with image blending. This paper proposes a new scheme that significantly reduce the unnatural edges and the color distortion. First, to detect and handle the incompletely cropped region, an adaptive binary line mask generation utilizing color difference checking algorithm (CDC) is proposed. The generated mask is exploited to im-prove image blending performance by isolating incompletely cropped image edges from image blending. Second, in order to perform inpaint-ing the missing or masked area of the object image and image blending together, the inpainting generative adversarial model is adopted. Exper-imental results show that the blended images are not only more natural than those of the previous works but the color information is also well preserved.

3D 영상 효과를 위한 레이어 채널 이미지의 처리 기법

최학현(Hak-Hyun Choi),김정희(Jung-Hee Kim),이명학(Myung-Hak Lee) 한국콘텐츠학회 2008 한국콘텐츠학회논문지 Vol.8 No.1

3D 영상에 이펙트를 표현할 수 있는 레이어 채널을 삽입함으로써 애플리케이션 렌더링에 효과적으로 이용하도록 한다. 현재의 이펙트 렌더링은 영상과 이펙트의 개별 처리 및 혼합의 방식을 사용하기 때문에 저장 공간과 영상 처리에 있어서 개별 소스를 필요로 하고 있다. 그러나 영상과 레이어 채널을 하나로 묶어 처리함으로써 비용 절약과 영상 처리 면에서 큰 효과를 볼 수 있다. 개발은 영상에 레이어 채널을 삽입하기 위해서 영상 포맷의 변경, 레이어 채널이 나타나지 않도록 숨김 기능 추가, 영상 로드시 영상과 레이어 채널을 동시 접근 가능하도록 제어, 영상과 레이어 채널이 쉽게 혼합될 수 있도록 간편한 알파블렌딩 처리 등의 방법으로 영상 포맷을 변경하여 레이어 채널을 숨기는 기법, 일반 영상 뷰어에서도 변경된 포맷의 영상을 볼 수 있도록 개발, 레이어 채널과 영상을 같이 묶음으로써 재사용성을 높이고 모든 프로그램에 이용 가능하도록 만든다. 그러면 영상 로드시에 영상과 레이어 채널을 동시에 불러드림으로써 처리 속도 향상시키고 3D 영상에 레이어 채널을 삽입함으로써 레이어 채널 영상을 위한 소스 저장 공간을 줄일 수 있다. 또한 3D 영상과 레이어 채널의 영상을 한 번에 다룰 수 있게 되어 효과적인 이펙트 표현 가능하고 실제 애플리케이션이 될 수 있는 멀티미디어 영상 등에 효과적으로 이용이 가능할 수 있을 것으로 기대한다. A layer channel, which can express effects on 3D image, is inserted to use it on application rendering effectively. The current method of effect rendering requires individual sources in storage and image processing, because it uses individual and mixed management of images and effects. However, we can save costs and improve results in images processing by processing both image and layer channels together. By changing image format to insert a layer channel in image and adding a hide function to conceal the layer channel and control to make it possible to approach image and layer channels simultaneously during loading image and techniques hiding the layer channel by changing image format with simple techniques, like alpha blending, etc., it is developed to improve reusability and be able to be used in all programs by combining the layer channel and image together, so that images in changed format can be viewed in general image viewers. With the configuration, we can improve processing speed by introducing image and layer channels simultaneously during loading images, and reduce the size of source storage space for layer channel images by inserting a layer channel in 3D images. Also, it allows managing images in 3D image and layer channels simultaneously, enabling effective expressions, and we can expect to use it effectively in multimedia image used in practical applications.

An X-ray Image Panorama System Using Robust Feature Matching and Per ception-Based Image Enhancement

Wang, Weiwei,Gwun, Oubong Korea Multimedia Society 2012 멀티미디어학회논문지 Vol.15 No.5

This paper presents an x-ray medical image panorama system which can overcome the smallness of the images that exist on a source computer during remote medical processing. In the system, after the standard medical image format DICOM is converted to the PC standard image format, a MSR algorithm is used to enhance X-ray images of low quality. Then SURF and Multi-band blending are applied to generate a panoramic image. Also, this paper evaluates the proposed SURF based system through the average gray value error and image quality criterion with X-ray image data by comparing with a SIFT based system. The results show that the proposed system is superior to SIFT based system in image quality.

모섬유의 혼방비율과 직물 특성에 따른 남성 정장용 소재의 질감이미지와 선호도 평가

김희숙(Kim, Heesook),나미희(Na, Mihee) 한국생활과학회 2011 한국생활과학회지 Vol.20 No.2

This research was designed to compare the subjective evaluation of texture image and preference according to fiber blending ratio of mens suit fabrics. 110 subjects evaluated the texture image and preference of various fabrics. For statistical analysis, factor analysis, MDS, pearson correlation and ANOVA were used. The results were as follows: Sensory image factors of suit fabrics were smoothness, bulkiness, stiffness, elasticity, moistness and weight sensation. Sensibility image factors were classic, practical, characteristic and sophisticated. Bulkiness and elasticity sensory images showed high correlations with sensibility images. Fabrics with high wool blending ratio showed as classic and sophisticated , bulkiness and elasticity texture images and fabrics with low wool blending ratio showed texture images of characteristic, surface character, stiffness, moistness and weight sensation. Wool fiber blending ratio affected on the purchase preference and tactile preference. Using regression analysis, it was shown that sensibility images had more of an effect on preference than sensory images. The thickness and pattern type showed positive effects and fiber blending ratio showed negative effects on the preference.

로컬 히스토그램 명세화에 기반한 화질 개선

울럭벡 쿠사노브,이창훈 한국지능시스템학회 2013 한국지능시스템학회논문지 Vol.23 No.1

In this paper we propose an image enhancement technique based on histogram specification method over local overlapping regions referred as Local Histogram Specification. First, both reference and original images are splitted into local regions that each overlaps half of its adjacent regions and general histogram specification method is used between corresponding local regions of reference and original image. However it produces noticeable boundary effects. Linear weighted image blending method is used to reduce this effect in order to make seamless image and we also proposed new technique dealing with over-enhanced contrast areas. We satisfied with our experimental results that showed better enhancement accuracy and less noise amplifications compared to other well-known image enhancement methods. We conclude that the proposed method is well suited for motion detection systems as a responsible part to overcome sudden illumination changes.

Image Enhancement Based on Local Histogram Specification

Ulugbek Khusanov(울럭벡 쿠사노브),Chang-Hoon Lee(이창훈) 한국지능시스템학회 2013 한국지능시스템학회논문지 Vol.23 No.1

In this paper we propose an image enhancement technique based on histogram specification method over local overlapping regions referred as Local Histogram Specification. First, both reference and original images are splitted into local regions that each overlaps half of its adjacent regions and general histogram specification method is used between corresponding local regions of reference and original image. However it produces noticeable boundary effects. Linear weighted image blending method is used to reduce this effect in order to make seamless image and we also proposed new technique dealing with over-enhanced contrast areas. We satisfied with our experimental results that showed better enhancement accuracy and less noise amplifications compared to other well-known image enhancement methods. We conclude that the proposed method is well suited for motion detection systems as a responsible part to overcome sudden illumination changes.

원격 제어 기반의 파노라마 이미지 생성 알고리즘

권오성 한국정보교육학회 2018 정보교육학회 학술논문집 Vol.9 No.1

본 논문에서는 원격지에서 듀얼 카메라가 연속 촬영한 복수 이미지를 수집하고 실시간 파노라마 형식으로 변 환하는 방식을 제안하고 구현하였다. 이를 위하여 원격 제어와 실시간 파노라마 이미지 생성 프로그램을 구현하 였다. 이미지 스티칭 구현은 일반적으로 널리 알려진 OpenCV 라이브러리를 사용하였고 적절한 입력과 처리를 위한 인터페이스와 변환 절차를 마련하여 실험하였다. 실험 결과, 적절한 변환과 예외 처리가 가능하였고 성공 적인 파노라마 변환 결과를 얻을 수 있었다. In this paper, we propose and implement an panorama generation algorithm for the stitching of real-time successive images. To implement our proposed method, we design an remote mechanism for camera shooting conrol and image stitching algorithm based on open source libraries. We make use of OpenCV libraries for implementing of our image stitching algorithm. In experimental, we can confirm an appropriate conversion and exception handling and take a successful panorama results.

An Analysis of Land Turtle Imagery in The Grapes of Wrath from the Perspective of Conceptual Blending

Xie Bin,Joseph Yosup Kim 한국영미문학교육학회 2020 영미문학교육 Vol.24 No.1

This study aims to analyze the cognitive process of land turtle image in Steinbeck’s The Grapes of Wrath based on the conceptual blending. Conceptual blending is a universal cognitive process in people’s thinking and activities, especially in creative thinking and actions. We have analyzed the physical features and crawling process of the turtle and the Joad family and constructed the corresponding relationship between these two. Using the spatial combination of conceptual blending and the completion of the background knowledge framework, the structure in the synthetic space was elaborated to form new meanings. Turtle, as the most prominent animal image in The Grapes of Wrath, bears rich symbolic connotation. Interpretation from the perspective of conceptual blending is conducive to readers to more precisely grasp the deep meaning hidden underneath the text, to attain a comprehensive understanding of the novel, to appreciate Steinbeck’s exquisite ideas and to realize the aesthetic experience of great works.

View synthesis with sparse light field for 6DoF immersive video

곽상운,윤정일,정준영,김영욱,임인성,정원식,서정일 한국전자통신연구원 2022 ETRI Journal Vol.44 No.1

Virtual view synthesis, which generates novel views similar to the characteristics of actually acquired images, is an essential technical component for delivering an immersive video with realistic binocular disparity and smooth motion parallax. This is typically achieved in sequence by warping the given images to the designated viewing position, blending warped images, and filling the remaining holes. When considering 6DoF use cases with huge motion, the warping method in patch unit is more preferable than other conventional methods running in pixel unit. Regarding the prior case, the quality of synthesized image is highly relevant to the means of blending. Based on such aspect, we proposed a novel blending architecture that exploits the similarity of the directions of rays and the distribution of depth values. By further employing the proposed method, results showed that more enhanced view was synthesized compared with the well-designed synthesizers used within moving picture expert group (MPEG-I). Moreover, we explained the GPU-based implementation synthesizing and rendering views in the level of real time by considering the applicability for immersive video service.

개념적 혼성에서 환유의 작용 방식에 관한 연구

김동환(Kim, Dong-Hwan) 미래영어영문학회 2017 영어영문학 Vol.22 No.1

The purpose of this paper is to deal with the issue of whether it is possible to incorporate metonymy into the framework of conceptual blending theory. It is usually claimed in cognitive linguistics that metonymy is a cognitive process in which one element provides mental access to another element in the same ICM (Idealized Cognitive Model). And conceptual blending theory posits the four-space network consisting of two input spaces, the generic space, and the blended space. At first glance, metonymy seems to be incompatible with conceptual blending theory. Here I will show two ways for metonymy to operate in conceptual blending. The first is the way metonymy operates within one input space, and the second is the way metonymy operates between two input spaces. Especially, for the second way, I will show the first discussion of metonymy in terms of conceptual blending theory by Coulson & Oakley (2003) and the discussion of the relationship of metonymy and conceptual blending at the neural level by Bierwiaczonek (2013).