• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.7
• /
• pp.2585-2598
• /
• 2015

Hand posture recognition has played a very important role in Human Computer Interaction (HCI) and Computer Vision (CV) for many years. The challenge arises mainly due to self-occlusions caused by the limited view of the camera. In this paper, a robust hand posture recognition approach based on 3D point cloud from two RGB-D sensors (Kinect) is proposed to make maximum use of 3D information from depth map. Through noise reduction and registering two point sets obtained satisfactory from two views as we designed, a multi-viewed hand posture point cloud with most 3D information can be acquired. Moreover, we utilize the accurate reconstruction and classify each point cloud by directly matching the normalized point set with the templates of different classes from dataset, which can reduce the training time and calculation. Experimental results based on posture dataset captured by Kinect sensors (from digit 1 to 10) demonstrate the effectiveness of the proposed method.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.3
• /
• pp.13-19
• /
• 2023

카메라 센서의 한계로 인하여 촬영 장면에 따라 한 번의 촬영으로 모든 영역의 밝기가 적절하게 촬영되지 않는 경우가 존재한다. 이러한 센서의 한계는 하이 다이나믹 레인지 이미징 기술을 통해서 극복이 가능하다. 한 장면을 다양한 노출 설정으로 여러 번 촬영하는 브라케팅은 움직이는 피사체를 찍기에 적절하지 않으며 촬영 시간이 길다는 단점이 있다. 본 연구는 한 번의 촬영으로 서로 다른 노출의 이미지를 얻을 수 있는 소형 라이트필드 카메라를 제안한다. 라이트필드 카메라는 대표적으로 두 가지 형태가 있는데, 첫 번째는 여러 대의 카메라를 어레이로 배치한 라이트필드 카메라 시스템이며, 두 번째는 대물렌즈 뒤에 마이크로 렌즈 어레이를 배치한 카메라이다. 본 연구에서 제작된 초박형 라이트필드 카메라는 센서 위에 마이크로 렌즈어레이가 부착되어있는 형태의 카메라로 각 렌즈 조리개 크기를 다르게 설계하여 한 번의 촬영으로 다른 노출의 촬영 결과를 얻을 수 있게 설계되었다. 촬영된 단일 영상들을 전처리 하여 이미지 품질을 높인 이후, HDR 알고리즘을 통해 각 단일 이미지들보다 다이나믹 레인지가 넓은 이미지를 획득하도록 구현하였다. 또한 노출 시간을 기준으로 설계된 식을 수정하여 조리개값에 따라 다른 가중치를 둘 수 있도록 바꾸었고, 이를 통해 단 한 번의 촬영을 통한 HDR 이미징을 구현하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.1
• /
• pp.35-41
• /
• 2015

When a video surveillance system tracks a specific object, it is very important to get quickly the information of the object through fast image processing. Usually one camera surveillance system for tracking the object made results in various problems such like occlusion, image noise during the tracking process. It makes difficulties on image based moving object tracking. Therefore, to overcome the difficulties the multi video surveillance system which installed several camera within interested area and looking the same object from multi angles of view could be considered as a solution. If multi cameras are used for tracking object, it is capable of making a decision having high accuracy in more wide space. This paper proposes a method of recognizing and tracking a specific object like a car using the homography in which multi cameras are installed at the crossroad.

• Journal of Broadcast Engineering
• /
• v.25 no.3
• /
• pp.439-448
• /
• 2020

In this paper, we propose a modified optimization algorithm for point cloud matching of multi-view RGB-D cameras. In general, in the computer vision field, it is very important to accurately estimate the position of the camera. The 3D model generation methods proposed in the previous research require a large number of cameras or expensive 3D cameras. Also, the methods of obtaining the external parameters of the camera through the 2D image have a large error. In this paper, we propose a matching technique for generating a 3D point cloud and mesh model that can provide omnidirectional free viewpoint using 8 low-cost RGB-D cameras. We propose a method that uses a depth map-based function optimization method with RGB images and obtains coordinate transformation parameters that can generate a high-quality 3D model without obtaining initial parameters.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.05a
• /
• pp.646-649
• /
• 2011

This paper presents the prototype for parking assistant system of multi-camera viewing. Proposed system provides the driver with three direction of the scenes surrounding vehicle. Depending on the handle rotation, the proposed system displays views from multiple wide-angle CCD cameras on the laptop screen to reduce blind spots, support smooth parallel or garage parking, and support comfortable and safe driving. The performance of roposed system is validated by vehicle experiments.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.2
• /
• pp.37-48
• /
• 2008

Generally, a motion vector and a disparity vector represent the motion information of an object in a single-view of camera and the displacement of the same scene between two cameras that located spatially different from each other, respectively. Conventional H.264/AVC does not use the disparity vector in the motion vector prediction because H.264/AVC has been developed for the single-view video. But, multi-view video coding that uses the inter-view prediction structure based on H.264/AVC can make use of the disparity vector instead of the motion vector when the current frame refers to the frame of different view. Therefore, in this paper, we propose an improved motion/disparity vector prediction method that consists of global disparity vector replacement and extended neighboring block prediction. From the experimental results of the proposed method compared with the conventional motion vector prediction of H.264/AVC, we achieved average 1.07% and 1.32% of BD (Bjontegaard delta)-bitrate saving for ${\pm}32$ and ${\pm}64$ of global vector search range, respectively, when the search range of the motion vector prediction is set to ${\pm}16$.

• Korean Journal of Remote Sensing
• /
• v.18 no.1
• /
• pp.53-59
• /
• 2002

The MSC is a payload on the KOMPSAT-2 satellite to perform the earth remote sensing. The instrument images the earth using a push-broom motion with a swath width of 15 km and a GSD(Ground Sample Distance) of 1 m over the entire FOV(Field Of View) at altitude 685 km. The instrument is designed to haute an on-orbit operation duty cycle of 20% over the mission lifetime of 3 years with the functions of programmable gain/offset and on-board image data compression/storage. The MSC instrument has one channel for panchromatic imaging and four channel for multi-spectral imaging covering the spectral range from 450nm to 900nm using TDI(Time Belayed Integration) CCD(Charge Coupled Device) FPA(Focal Plane Assembly). The MSC hardware consists of three subsystem, EOS(Electro Optic camera Subsystem), PMU(Payload Management Unit) and PDTS(Payload Data Transmission Subsystem) and each subsystems are currently under development and will be integrated and verified through functional and space environment tests. Final verified MSC will be delivered to spacecraft bus for AIT(Assembly, Integration and Test) and then COMSAT-2 satellite will be launched after verification process through IST(Integrated Satellite Test). In this paper, the introduction of MSC, the configuration of MSC electronics including electrical interlace and design of CEU(Camera Electronic Unit) in EOS are described. MSC Operation parameters induced from the operation concept are discussed and analyzed to find the influence of system for on-orbit operation in future.

• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.313-321
• /
• 2017

One of the most frequently performed tasks in human-robot interaction (HRI), intelligent vehicles, and security systems is face related applications such as face recognition, facial expression recognition, driver state monitoring, and gaze estimation. In these applications, accurate head pose estimation is an important issue. However, conventional methods have been lacking in accuracy, robustness or processing speed in practical use. In this paper, we propose a novel method for estimating head pose with a monocular camera. The proposed algorithm is based on a deep neural network for multi-task learning using a small grayscale image. This network jointly detects multi-view faces and estimates head pose in hard environmental conditions such as illumination change and large pose change. The proposed framework quantitatively and qualitatively outperforms the state-of-the-art method with an average head pose mean error of less than $4.5^{\circ}$ in real-time.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.10
• /
• pp.4117-4135
• /
• 2020

Due to the angle of view, illumination and scene diversity, real-time dynamic face detection and recognition is no small difficulty in those unrestricted environments. In this study, we used the intrinsic correlation between detection and calibration, using a multi-task cascaded convolutional neural network(MTCNN) to improve the efficiency of face recognition, and the output of each core network is mapped in parallel to a compact Euclidean space, where distance represents the similarity of facial features, so that the target face can be identified as quickly as possible, without waiting for all network iteration calculations to complete the recognition results. And after the angle of the target face and the illumination change, the correlation between the recognition results can be well obtained. In the actual application scenario, we use a multi-camera real-time monitoring system to perform face matching and recognition using successive frames acquired from different angles. The effectiveness of the method was verified by several real-time monitoring experiments, and good results were obtained.

• International Journal of Internet, Broadcasting and Communication
• /
• v.7 no.1
• /
• pp.15-19
• /
• 2015

The depth information of an image is used in a variety of applications including 2D/3D conversion, multi-view extraction, modeling, depth keying, etc. There are various methods to acquire depth information, such as the method to use a stereo camera, the method to use the depth camera of flight time (TOF) method, the method to use 3D modeling software, the method to use 3D scanner and the method to use a structured light just like Microsoft's Kinect. In particular, the depth camera of TOF method measures the distance using infrared light, whereas TOF sensor depends on the sensitivity of optical light of an image sensor (CCD/CMOS). Thus, it is mandatory for the existing image sensors to get an infrared light image by bundling several pixels; these requirements generate a phenomenon to reduce the resolution of an image. This thesis proposed a measure to acquire a high-resolution image through gradual area movement while acquiring a low-resolution image through pixel bundling method. From this measure, one can obtain an effect of acquiring image information in which illumination intensity (lux) and resolution were improved without increasing the performance of an image sensor since the image resolution is not improved as resolving a low-illumination intensity (lux) in accordance with the gradual pixel bundling algorithm.