• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2008년도 춘계종합학술대회 A
• /
• pp.585-588
• /
• 2008

본 논문은 손동작 인식을 위한 형태 유사도 측정 방법을 제안한다. 이를 위해 손 영역 획득과 유사도 측정 단계로 나눈다. 손 영역 획득은 YCbCr 칼라 공간을 이용하여 손 영역을 추출하며, filter와 Histogram분석을 통하여 노이즈를 제거한다. 그리고 손 형태 유사도 측정은 윤곽선을 추출한 후 인접 간선들 사이의 거리와 각도 관계로 TSR을 적용하여 손동작의 유사성을 측정하였다. 파악된 특징점으로부터 형태 유사도 값을 측정한 후, 이를 손동작을 인식하는데 활용한다.

• 한국멀티미디어학회논문지
• /
• 제19권4호
• /
• pp.717-724
• /
• 2016

Paintings can evoke emotions in viewers. In this paper, we propose a method for extracting emotion from paintings by using the colors that comprise the paintings. For this, we generate color spectrum from input painting and compare the color spectrum and color combination for finding most similarity color combination. The found color combinations are mapped with emotional keywords. Thus, we extract emotional keyword as the emotion evoked by the painting. Also, we vary the form of algorithms for matching color spectrum and color combinations and extract and compare results by using each algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권3호
• /
• pp.1121-1141
• /
• 2020

We propose a robust visual object tracking algorithm fusing a convolutional neural network tracker trained offline from a large number of video repositories and a color histogram based tracker to track objects for mixing immersive audio. Our algorithm addresses the problem of occlusion and large movements of the CNN based GOTURN generic object tracker. The key idea is the offline training of a binary classifier with the color histogram similarity values estimated via both trackers used in this method to opt appropriate tracker for target tracking and update both trackers with the predicted bounding box position of the target to continue tracking. Furthermore, a histogram similarity constraint is applied before updating the trackers to maximize the tracking accuracy. Finally, we compute the depth(z) of the target object by one of the prominent unsupervised monocular depth estimation algorithms to ensure the necessary 3D position of the tracked object to mix the immersive audio into that object. Our proposed algorithm demonstrates about 2% improved accuracy over the outperforming GOTURN algorithm in the existing VOT2014 tracking benchmark. Additionally, our tracker also works well to track multiple objects utilizing the concept of single object tracker but no demonstrations on any MOT benchmark.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제36권1호
• /
• pp.54-61
• /
• 2009

피부색 정보는 비전 기반 시스템에서 인체 인식에 널리 쓰이는 중요한 정보이다. 그러나 기존의 픽셀 단위의 피부색 분할 방법은 피부색 영역 내부와 외부에 발생하는 오분할로 인해 여러 가지 피부색 관련 시스템의 인식률을 저해시키는 요인이 된다. 본 논문에서는 양자화 영역 정보로부터 프레임 간에 근접한 유사 피부색의 영역별 분할을 통한 피부색 분할 방법을 제안한다. 제안하는 방법은 피부색 영역분할을 위해 JSEG 알고리즘을 통해 영상의 칼라를 양자화하여 영역을 분할한다. 분할된 영역으로부터 근접한 유사 피부 영역의 후보를 결정하고, 각 영역의 히스토그램 비교를 통해 피부색 영역을 결정한다. 이렇게 결정된 영역으로부터 피부색 표본을 추출하여 다음 프레임을 위한 피부색 모델을 갱신한다. 성능 평가를 위해 ECHO 데이타베이스와 조명이 변화하는 환경에서 실제 촬영한 영상을 이용하여 기존 연구의 분류 방법 비교 실험을 실시하였고, 기존보다 향상된 영역 분할 및 조명 적응 성능을 보였다.

• 복식문화연구
• /
• 제25권6호
• /
• pp.819-830
• /
• 2017

This study aims to gather precise information on the real fabric color and texture, and purchasing intention of mobile shoppers buying clothes. Eighty volunteers participated in the sensory test on three smartphones with four colors and two fabrics-smooth taffeta and hairy doeskin. This study carried out the posteriori test using the one-way ANOVA and Duncan test by SPSS21.0. In the analysis' results of color preference, there were no differences among the four colors of taffeta between the smartphones, but different preferences between the red and yellow doeskin exist. In the case of the Samsung phone, which has an immense color distortion, the red fabric has a low color preference. In contrast, on the Apple phone yellow fabric had the highest preference because of its brightness. The Apple phone also has the highest purchasing intention of yellow colored taffeta, which is similar to the color preference results, although the real fabric has the opposite result. For doeskin, the real red and blue colored fabric has the highest purchasing intention. The Samsung phone has the biggest color mismatch with the real fabric. It also has the lowest purchasing intention of red taffeta fabric, while the LG phone has the lowest purchasing intention of blue fabric. Using the paired comparison method of the similarity between 'real' fabrics and the mobile version of fabric colors has a low similarity on all four colors of taffeta and doeskin fabrics. Therefore it can be concluded that phones do not represent the 'real' fabric color.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 1996년도 학술대회
• /
• pp.63-66
• /
• 1996

In this paper, we propose a new content-based color image retrieval algorithm. The algorithm makes use of two features; colors as global features and block classification results as local features. More specifically, we obtain R, G, B color histograms and classify nonoverlapping small image blocks into texture, monotone, and various edges, then using these histograms and classification results were make a similarity measure. Experimental results show that retrieval rate of the proposed algorithm is higher than the previous method.

• 전자공학회논문지S
• /
• 제34S권12호
• /
• pp.81-89
• /
• 1997

In this paper, we propose a new image retrieval algorithm using the block classification. More specifically, we classify nonoverlappint small image blocks into texture, monotone, and various edges. Using these classification results and the RGB color histogram, we propose a new similarity measure which considers both local and global fretures. According to our experimental results using 232 color images, the retrieval efficiencies of the proposed and the previous methods were 0.610 and 0.522, respectively, which implies that the proposed algorithm yields better performance.

• 감성과학
• /
• 제7권4호
• /
• pp.57-65
• /
• 2004

본 논문에서는 color, texture, shape의 정보를 통합 이용하여 내용기반 영상검색 시스템의 성능을 향상시키는 기법을 고찰하였다. 먼저 영상에 내재되어 있는 color를 분석 추출하여 몇 개의 대표색으로 요약 표현한 다음, 이를 활용한 근사치 측정도를 고안하였다. Texture정보 분석에 있어서는 영상의 주축 행렬 데이터를 통계적 접근 방법으로 추출하였다. Edge분석의 방법으로는 Edge 막대그래프에서 색상변환, 양자화, 필터링에 관련된 정보를 선행처리 후 Edge 정보를 추출하였다. 마지막으로, 본 연구의 결과인 내용기반 영상검색 시스템의 효율성을 precision-recall 분석과 실험적 결과를 통하여 입증하였다.

• 복식
• /
• 제58권4호
• /
• pp.35-46
• /
• 2008

The purpose of this study was to closely examine the history of glasses and their cases used in the middle and late of Joseon Dynasty and identify inherent quality of our traditional glasses cases through color analyses of glasses cases' material and shape. While theoretical examination was conducted based on the literatures of glasses and their cases that firstly appeared in around Japanese Invasion (Imjin war) of Korea in 1592, practical analyses were demonstrated on photos of glasses cases used in the middle and late of Joseon Dynasty collected from both museum pieces and the internet which were grouped into wood, fabric, paper, sharkskin, hawksbill and cow's horn in accordance with their materials. 623 color samples were abstracted from collected 159 glasses cases and quantity analyses on each material were performed respectively. Abstracted representative colors based on the result of color analyses were classified into the main materials and accessories' color scheme. The result of this study are as follow: firstly, both Yellow and Yellow Red were mostly used in main materials. In Fabric case's colors were widely used in embroidery and in animal matter material cases such as sharkskin, hawksbill and cow's horn, which can be used as itself or dyed, Green Yellow shown in high frequency. Secondly, accessories were analyzed into similarity coloration with main materials. From this finding, it turns out that our traditional cases have characteristic of similarity coloration between main materials and accessories. Red Purple and Purple Blue in high frequency in accessories used as an accent color. Finally, based on the analysis of hue and tone, while the middle and low value colors shown in very high frequency, the high-chroma colors hardly shown.

• 컴퓨터교육학회논문지
• /
• 제7권5호
• /
• pp.111-120
• /
• 2004

영상의 칼라, 텍스쳐, 오브젝트의 형체 등과 같은 하위 수준의 특징을 표현할 수 있는 기술자를 MPEG-7 표준에서 규정하고 있다. 하지만, 각각의 기술자를 따로 분석함으로써는 성능 향상에 불충분한 점이 있었다. 본 논문에서는 칼라 기술자와 텍스쳐 기술자를 결합하여 영상검색의 성능을 향상시키는 방법을 제안한다. MPEG-7 표준에서 정의한 $l_{1}$-norm방법보다, 본 논문에서는 칼라 히스토그램의 경우 코사인 근사도 계수를, 에지 히스토그램의 경우 유클리디언 디스턴스를 적용 실험하여 진일보한 결과를 도출할 수 있었다.