• 방송공학회논문지
• /
• 제23권3호
• /
• pp.361-368
• /
• 2018

자율주행 자동차의 연구가 빠르게 발전하는 가운데 보행자 검출에 대한 연구 또한 성공적으로 진행되고 있다. 그러나 대부분의 연구에서 사용되는 데이터셋이 컬러영상을 기반하고 있고 또한 보행자의 인식이 상대적으로 쉬운 영상이 많다. 컬러 영상의 경우 보행자가 빛에 노출되는 정도에 따라 영상에 제대로 포착이 되지 않을 수 있고 이로 인해 기존 방식들로는 이러한 보행자를 제대로 검출하지 못하는 상황이 발생한다. 따라서 본 논문에서는 DNN (deep neural network) 기반 컬러 영상과 열 영상을 이용한 다중 스펙트럼 보행자 검출 기법을 제안하고자 한다. 기존의 SSD (single shot multibox detector) 기법을 기반으로 하여 컬러 영상과 열 영상을 동시에 활용하는 퓨전 네트워크 구조를 제안한다. 실험은 KAIST의 데이터셋을 이용하여 실시하였고 제안한 기법인 SSD-H (SSD-Halfway fusion)의 방식이 KAIST 보행자 검출기준의 기준치보다 18.18% 낮은 miss rate를 획득하였고 또한 기존 halfway fusion 기법에 비해 최소 2.1% 낮은 miss rate를 획득하였다.

• 방송공학회논문지
• /
• 제16권3호
• /
• pp.561-565
• /
• 2011

본 논문에서는 기존의 베이어 컬러필터 배열 기반의 영상 센서를 사용하는 카메라에서 개선된 컬러 영상을 생성하기 위한 영상 시스템을 제안한다. 제안한 컬러 영상 생성 기술은 두 장의 연속적인 영상(광대역 G채널 필터를 사용한 영상과 색필터를 적용하지 않고 연속적으로 촬영한 영상)으로부터 RGBCY의 다양한 색 정보를 추출하여 디모자이킹을 위한 컬러 영상 보간의 계산은 감소시키며 개선된 색상 정보를 가지는 영상을 얻을 수 있다. 또한 기존의 베이어 CFA 영상센서를 사용하기 때문에 새로운 형태의 CFA를 제작할 필요 없이 다중 스펙트럼 영상을 취득할 수 있는 카메라 시스템으로 확장 가능하다. 마지막으로 본 논문에서 제안한 컬러영상 생성을 위한 카메라 시스템의 성능을 검증하기 위해서 기존의 디모자이킹 기술과 비교하였으며, 그 결과 컬러 중첩 현상이 상당히 개선되었음을 보였다.

• Restorative Dentistry and Endodontics
• /
• 제48권3호
• /
• pp.26.1-26.12
• /
• 2023

Objective: This in vitro study aimed to evaluate the effects of different whitening toothpastes on a composite resin during at-home bleaching with 10% carbamide peroxide. Materials and Methods: Sixty samples (7 mm × 2 mm) were used for color and roughness analyses, while another 60 samples (3 mm × 2 mm) were utilized to assess microhardness. The factors analyzed included toothpaste, for which 5 options with varying active agents were tested (distilled water; conventional toothpaste; whitening toothpaste with abrasive agents; whitening toothpaste with abrasive and chemical agents; and whitening toothpaste with abrasive, chemical, and bleaching agents). Brushing and application of whitening gel were performed for 14 days. Surface microhardness (SMH), surface roughness (Ra), and color (ΔL^*, Δa^*, Δb, ΔE^*_ab, and ΔE₀₀) were analyzed. The Ra and SMH data were analyzed using mixed generalized linear models for repeated measures, while the color results were assessed using the Kruskal-Wallis and Dunn tests. Results: Between the initial and final time points, all groups demonstrated significant increases in Ra and reductions in SMH. No significant differences were found between groups for SMH at the final time point, at which all groups differed from the distilled water group. Conventional toothpaste exhibited the lowest Ra, while whitening toothpaste with abrasive agent had the highest value. No significant differences were observed in ΔL^*, Δa^*, and Δb. Conclusions: While toothpaste composition did not affect the color stability and microhardness of resin composite, combining toothbrushing with whitening toothpaste and at-home bleaching enhanced the change in Ra.

• 한국축산식품학회지
• /
• 제44권2호
• /
• pp.356-371
• /
• 2024

Novel meat-inspired products, such as cell-cultivated meat and meat analogues, embrace environmental sustainability, food safety and security, animal welfare, and human health, but consumers are still hesitant to accept these products. The appearance of food is often the most persuasive determinant of purchasing decisions for food. Producing cultivated meat and meat analogues with similar characteristics to conventional meat could lead to increased acceptability, marketability, and profitability. Color is one of the sensorial characteristics that can be improved using color-inducing methods and colorants. Synthetic colorants are cheap and stable, but natural pigments are regarded as safer components for novel food production. The complexity of identifying specific colorants to imitate both raw and cooked meat color lies in the differences in ingredients and methods used to produce meat alternatives. Research devoted to improving the sensorial characteristics of meat analogues has noted various color-inducing methods (e.g., ohmic cooking and pasteurization) and additives (e.g., lactoferrin, laccase, xylose, and pectin). Additionally, considerations toward other meat components, such as fat, can aid in mimicking conventional meat appearance. For instance, the use of plant-based fat replacers and scaffolds can produce a marked sensory enhancement without compromising the sustainability of alternative meats. Moving forward, consumer-relevant sensorial characteristics, such as taste and texture, should be prioritized alongside improving the coloration of meat alternatives.

• 대한전자공학회논문지SP
• /
• 제42권1호
• /
• pp.25-31
• /
• 2005

본 논문에서는 컬러 영상에서 Support Vector Domain Description (SVDD)를 이용한 얼굴 검출 방법을 제안한다. 기존의 훈련을 통한 얼굴 검출 방법은 얼굴 영상과 얼굴이 아닌 영상을 모두 사용해야 한다. 그러나, SVDD를 이용한 얼굴 검출은 단지 훈련을 위해 얼굴 영상만이 사용된다. SVDD의 훈련을 통해 나오는 값인 반지름과 중심 좌표를 통해 얼굴을 검출한다. 또한, 엔트로피를 이용한 임계값 추출 방법(Entropic Threshold)을 통해 얼굴 특징을 추출하고, 슬라이딩 윈도우(sliding window)기법을 통해 성능을 개선한다. 주성분 분석(Principle Component Analysis) 과 SVDD를 이용한 얼굴 검출 방법의 비교 실험을 통해 본 논문이 제안한 방법의 효율성을 확인한다.

• 복식
• /
• 제54권3호
• /
• pp.13-25
• /
• 2004

The purpose of this study was to show the analysis system and the expression which is applied to fashion illustration by three major components in visual sign, metaphor and metonymy. The results of this study were as follows : Firstly, metaphor in qualisign of syntactics was revealed as a color such as realistic description, a pattern such as clothing of figure. etc. Metonymy was revealed as a social and cultural background color, a concept pattern. etc. In sinsign of syntactics. metaphor was revealed as a human body, non-human body and metamorphosis human body and metonymy as a human body and non-human body. In legisign of syntactics, the metaphor by perspective was used for a fantasy of space. and the metonymy was revealed as a color perspective representation, etc. The degree of change of texture was revealed as a metaphor and metonymy of gradation. And conventional custom sign was almost revealed in metaphor. Secondly, semantics showed about fashion image as juxtaposition of two similar objects in metaphor and as real description and simplification in metonymy Alternative fashion image in semantics was presented as a object related to fashion image. Conventional symbolic sign was presented as a role to clarify a fashion message in metaphor. Thirdly, the metaphorical and metonymical expression in pragmatics were usually presented as drawing and painting.

• 한국멀티미디어학회논문지
• /
• 제13권11호
• /
• pp.1621-1629
• /
• 2010

본 논문에서는 칼라비전의 색사상에서 가장 많이 활용되는 MSR(multi-scale Retinex) 기법의 속도를 크게 개선한 MSRCR(MSR with color restoration) 알고리즘을 제시한다. 기존 MSR기법은 보통 3개의 SSR(single-scale Retinex)로 구성되며 각 SSR에 크기가 다른 Gaussian 주변함수를 사용하고 있으며, 이 함수와의 상승적분 부분에서 많은 계산이 요구된다. 그러므로 제안한 알고리즘은 속도를 높이기 위해 Gaussian 함수와 등가적인 HDC(hierarchical discrete correlation)를 사용하고 휘도영상에만 적용하는 기법을 제시하며, 휘도영상의 Retinex 결과 값을 이용하여 색이 보존되는 단순한 MSRCR 알고리즘을 개발하였다. 실험을 통하여 제안한 기법은 기존의 가장 단순한 MSR기법보다 연산량 및 속도를 1/9.5배, 1/3.5배로 줄일 수 있었으며 기존 기법과 동등한 결과를 얻을 수 있었다.

• 한국멀티미디어학회논문지
• /
• 제18권9호
• /
• pp.1027-1038
• /
• 2015

For last several decades, the color correction methods have been proposed for HDR(high dynamic range) images. However, color distortion problems take place after correcting the colors such as halos, dominant color as well known. Accordingly, this article presents a novel approach in which the method consists of tone-mapping method and cone response function. In the proposed method, the tone mapping method is used to improve the contrast in the given HDR image based on chromatic and achromatic based on the CIEXYZ tristimulus value, expressed in c/m². The cone response function is used to deal with mismatch between corrected image and displayed image as well as to estimate various human visual effects based on the CMCAT2000 color appearance model. The experimental results show that the proposed method yields better performance of color correction over the conventional method in subjective and quantitative quality, and color reproduction.

• 한국재료학회지
• /
• 제28권4호
• /
• pp.221-226
• /
• 2018

The property changes of 18, 14, and 8K green gold alloys for jewelry are observed by adding 0.0, 3.0, and 5.0 wt% of indium (In), respectively. To check the composition of the alloys, an energy dispersive spectroscopy (EDS) analysis is conducted. Color and microstructure analysis is executed through bare-eye, macro camera, UV-VIS-NIR-colormeter, and optical microscope. The melting point, wetting angle, and hardness are measured using TGA-DTA, a wetting angle tester, and a Vickers hardness tester. The EDS analysis result demonstrates that each of the green gold alloys was manufactured with purposed contents. The color analysis result shows that the color of the alloys is similar to the color of the conventional 4 wt%-Cd 18K green gold, and the green color improves as the In content increases. The micro structure analysis result demonstrates that grain refinement improves as the amount of In increases. Enhancements in the melting point, wettability, and Vickers hardness changes appear as the In content increases and Au content decreases. The hardness is up to 260, which implies good durability. Therefore, the results suggest that the proposed 18, 14, and 8K In-added green gold alloys enhance the properties of jewelry products with regard to the green color, castability, and durability.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2004년도 춘계학술대회 학술발표 논문집 제14권 제1호
• /
• pp.201-205
• /
• 2004

The digital color image can be distorted by noise for a transmission or other elements of system. It happens to vague of a boundary side in the division of a color image object, especially, boundary side of an input color image is very important because it can be determined to the division and detection element in pattern recognition. Therefore it is boundary part In this paper, it detects the optimal edge with applying this color image to WCNN algorithm, after it does level up a boundary side of a color image by using the adaptive morphology as the threshold of an input color image. Also, it is used not a conventional fixed mask edge detection method but variable mask method which is cal led a variable BBM. It is confirmed by simulation that the proposed algorithm can be got the batter result edge at the place of closing to each edges and having smoothly curved line.