• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04b
• /
• pp.571-573
• /
• 2000

복잡한 이미지에서 얼굴 추출은 얼굴 영상처리 분야에서 기본적이면서도 배경이 복잡함으로 인해 많은 어려움이 따른다. 이 논문에서는 복잡한 화면 이미지에서 얼굴을 추출하기 위해 여러 가지 과정을 거친다. 다양한 피부색을 가진 얼굴에 대해 즉, 흑인과 황인, 백인 등을 모두 추출하기 위해 피부색 모델을 이용한다. 다양한 피부색에 대한 임계값(threshold)을 이용하여 피부색과 다른 영역을 구분하여 얼굴의 후보 데이터로 추출한다. 그 추출된 후보 데이터를 지역적 임계값(local threshold)을 이용하여 얼굴과 눈, 코, 입과 같은 세부사항에 분류한다. 분류된 부분이 즉 얼굴내에서 얼굴이 아닌 부분(눈, 코, 입 등)의 크기가 정규화 되어진 최소 크기보다 박을 경우 그 후보 데이터를 버리고, 그렇지 않을 경우, 즉 얼굴이 아닌 다른 부분의 크기가 정해진 크기보다 크거나 같을 경우 그 후보 데이터를 검출한다. 이 논문에 결과는 배경에서도 피부색과 얼굴의 부분영역을 이용하여 얼굴을 검출할 수 있다는 것을 보인다.

• Smart Media Journal
• /
• v.1 no.4
• /
• pp.27-34
• /
• 2012

In this paper, a novel text binarization is presented that can deal with some complex conditions, such as shadows, non-uniform illumination due to highlight or object projection, and messy backgrounds. To locate the target text region, a focus line is assumed to pass through a text region. Next, connected component analysis and stroke width estimation based on location information of the focus line is used to locate the bounding box of the text region, and each box of connected components. A series of classifications are applied to identify whether each CC(Connected component) is text or non-text. Also, a modified K-means clustering method based on an HCL color space is applied to reduce the color dimension. A text binarization procedure based on location of text component and seed color pixel is then used to generate the final result.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.941-942
• /
• 2006

A depth measurement system to recognize 3D shape of objects using single camera, line laser and a rotating mirror has been investigated. The camera and the light source are fixed, facing the rotating mirror. The laser light is reflected by the mirror and projected to the scene objects whose locations are to be determined. The camera detects the laser light location on object surfaces through the same mirror. The scan over the area to be measured is done by mirror rotation. The Segmentation process of object recognition is performed using the depth data of restored 3D data. The Object recognition domain can be reduced by separating area of interest objects from complex background.

• Journal of The Korean Digital Architecture Interior Association
• /
• v.12 no.2
• /
• pp.33-38
• /
• 2012

The purpose of luminance measurement in lighting design of urban scale is to prevent light trespass and light pollution, caused by over-bright. Specially, the illumination of road, tunnel and sports complex as urban infrastructure is measured by international standard for illuminance and luminance. It has been used the spot luminance meter and the image photometer to define the characteristics of qualitative and quantitative in exterior lighting. This study introduced the concept of HDR Image, procedure of HDR Image creation and analysis. Applications of HDR Image, particularly within the apartment building exterior lighting design analysis, have been demonstrated. The procedure proposed in this study is a affordable method that is useful for the assessment of luminance distribution in the scene with electric light. It also provides a measurement capability with the high resolution luminance data within a large field of view efficiently and quickly, which is not possible to achieve with a luminance meter.

• Journal of Sensor Science and Technology
• /
• v.22 no.6
• /
• pp.387-392
• /
• 2013

Camera rigs for shooting 3D video are classified as manual, motorized, or fully automatic. Even in an automatic camera rig, the process of Stereoscopic 3D (S3D) video capture is very complex and time-consuming. One of the key time-consuming operations is capturing the distance parameters, which are near distance, far distance, and convergence distance. Traditionally these distances are measured by tape measure or triangular indirect measurement methods. These two methods consume a long time for every scene in shot. In our study, a compact laser distance sensing system with long range distance sensitivity is developed. The system is small enough to be installed on top of a camera and the measuring accuracy is within 2% even at a range of 50 m. The shooting time of an automatic camera rig equipped with the laser distance sensing system can be reduced significantly to less than a minute.

• Korean Journal of Remote Sensing
• /
• v.22 no.5
• /
• pp.351-356
• /
• 2006

Based on current practical needs for geo-spatial information on mobile platform, the main theme of this study is a design and implementation of dynamic 3D terrain rendering system using spaceborne imagery, as a kind of texture image for photo-realistic 3D scene generation on mobile environment. Image processing and 3D graphic techniques and algorithms, such as TIN-based vertex generation with regular spacing elevation data for generating 3D terrain surface, image tiling and image-vertex texturing in order to resolve limited resource of mobile devices, were applied and implemented by using graphic pipeline of OpenGL|ES (Embedded System) API. Through this implementation and its tested results with actual data sets of DEM and satellite imagery, we demonstrated the realizable possibility and adaptation of complex typed and large sized 3D geo-spatial information in mobile devices. This prototype system can be used to mobile 3D applications with DEM and satellite imagery in near future.

• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.399-410
• /
• 2021

Aiming at the problem that complex external variables in natural scenes have a greater impact on facial expression recognition results, a facial expression recognition method based on two-stream convolutional neural network is proposed. The model introduces exponentially enhanced shared input weights before each level of convolution input, and uses soft attention mechanism modules on the space-time features of the combination of static and dynamic streams. This enables the network to autonomously find areas that are more relevant to the expression category and pay more attention to these areas. Through these means, the information of irrelevant interference areas is suppressed. In order to solve the problem of poor local robustness caused by lighting and expression changes, this paper also performs lighting preprocessing with the lighting preprocessing chain algorithm to eliminate most of the lighting effects. Experimental results on AFEW6.0 and Multi-PIE datasets show that the recognition rates of this method are 95.05% and 61.40%, respectively, which are better than other comparison methods.

• Journal of the Korean Society of Industry Convergence
• /
• v.26 no.5
• /
• pp.785-792
• /
• 2023

Because hyperspectral imaging provides detailed spectral information across a broad range of wavelengths, it can be utilized in numerous applications, including environmental monitoring, food quality inspection, medical diagnosis, material identification, art authentication, and crime scene analysis. However, hyperspectral images often contain various types of distortions due to the environmental conditions during image acquisition, which necessitates the proper removal of these distortions through a data preprocessing process. In this study, a preprocessing method was investigated to effectively correct the distortion caused by artificial light sources used in indoor hyperspectral imaging. For this purpose, a halogen-tungsten artificial light source was installed indoors, and hyperspectral images were acquired. The acquired images were then corrected for distortion using a preprocessing that does not require complex auxiliary equipment. After the corrections were made, the results were analyzed. According to the analysis, a statistical transformation technique using mean and standard deviation with reference to a reference signal was found to be the most effective in correcting distortions caused by artificial light sources.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.1
• /
• pp.41-48
• /
• 2009

This Paper introduces representation and coding schemes of multiview texture and depth data for complex three-dimensional scenes. We represent input color and depth images using compressed texture and depth map pairs. The proposed X-codec encodes them further to increase compression ratio in a near-lossless way. Our system resolves two problems. First, rendering time and output visual quality depend on input image resolutions rather than scene complexity since a depth image-based rendering techniques is used. Second, the random access problem of conventional image-based rendering could be effectively solved using our image block-based compression schemes. From experimental results, the proposed approach is useful to graphics applications because it provides multiview rendering, selective decoding, and scene manipulation functionalities.

• Journal of Information Processing Systems
• /
• v.14 no.2
• /
• pp.469-486
• /
• 2018

We propose an enhanced version of the local binary pattern (LBP) operator for texture extraction in images in the context of image retrieval. The novelty of our proposal is based on the observation that the LBP exploits only the lowest kind of local information through the global histogram. However, such global Histograms reflect only the statistical distribution of the various LBP codes in the image. The block based LBP, which uses local histograms of the LBP, was one of few tentative to catch higher level textural information. We believe that important local and useful information in between the two levels is just ignored by the two schemas. The newly developed method: gradual locality integration of binary patterns (GLIBP) is a novel attempt to catch as much local information as possible, in a gradual fashion. Indeed, GLIBP aggregates the texture features present in grayscale images extracted by LBP through a complex structure. The used framework is comprised of a multitude of ellipse-shaped regions that are arranged in circular-concentric forms of increasing size. The framework of ellipses is in fact derived from a simple parameterized generator. In addition, the elliptic forms allow targeting texture directionality, which is a very useful property in texture characterization. In addition, the general framework of ellipses allows for taking into account the spatial information (specifically rotation). The effectiveness of GLIBP was investigated on the Corel-1K (Wang) dataset. It was also compared to published works including the very effective DLEP. Results show significant higher or comparable performance of GLIBP with regard to the other methods, which qualifies it as a good tool for scene images retrieval.