• Journal of Korea Game Society
• /
• v.11 no.1
• /
• pp.147-157
• /
• 2011

Webcam doesn't perform memory-alignment in order to reduce the transmission time of image data. Memory-unaligned image data is unsuitable for the processing on GPU. Accordingly, we convert it to available color format for optimization in high speed image processing. In this paper, we propose a technique that accelerates webcam's color format conversion by using NVDIA CUDA. We propose an optimization which is about memory accesses and thread composition, also evaluate memory and computing performance for verifying a hypothesis which is the performance of the proposed architecture and optimizing degree on low-performance GPU. Following the optimization technique, we show performance improvements over maximum 68 percent.

• Journal of Digital Convergence
• /
• v.15 no.4
• /
• pp.267-273
• /
• 2017

With the development of various devices and communication technologies, the production and distribution of various multimedia contents are increasing exponentially. In order to retrieve multimedia data such as images and videos, an approach different from conventional text-based retrieval is needed. Color and shape are key features used in content-based image retrieval, which quantifies and analyzes various physical features of images and compares them to search for similar images. Color and shape have been used as independent features, but the two features are closely related in terms of cognition. In this paper, a method of describing the spatial distribution of color using a complex color model that projects three-dimensional color information onto two-dimensional complex form is proposed. Experimental results show that the proposed method can efficiently represent the shape of spatial distribution of colors by frequency transforming the complex image and reconstructing it with only a few coefficients in the low frequency.

• The KIPS Transactions:PartB
• /
• v.9B no.5
• /
• pp.563-570
• /
• 2002

Robust extraction of 3D facial features and global motion information from 2D image sequence for the MPEG-4 SNHC face model encoding is described. The facial regions are detected from image sequence using multi-modal fusion technique that combines range, color and motion information. 23 facial features among the MPEG-4 FDP (Face Definition Parameters) are extracted automatically inside the facial region using color transform (GSCD, BWCD) and morphological processing. The extracted facial features are used to recover the 3D shape and global motion of the object using paraperspective camera model and SVD (Singular Value Decomposition) factorization method. A 3D synthetic object is designed and tested to show the performance of proposed algorithm. The recovered 3D motion information is transformed into global motion parameters of FAP (Face Animation Parameters) of the MPEG-4 to synchronize a generic face model with a real face.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.26 no.2
• /
• pp.28-36
• /
• 2022

As port structures are exposed to various extreme external loads such as wind (typhoons), sea waves, or collision with ships; it is important to evaluate the structural safety periodically. To monitor the port structure, especially the rubber fender, a fender segmentation system using a vision sensor and deep learning method has been proposed in this study. For fender segmentation, a new deep learning network that improves the encoder-decoder framework with the receptive field block convolution module inspired by the eccentric function of the human visual system into the DenseNet format has been proposed. In order to train the network, various fender images such as BP, V, cell, cylindrical, and tire-types have been collected, and the images are augmented by applying four augmentation methods such as elastic distortion, horizontal flip, color jitter, and affine transforms. The proposed algorithm has been trained and verified with the collected various types of fender images, and the performance results showed that the system precisely segmented in real time with high IoU rate (84%) and F1 score (90%) in comparison with the conventional segmentation model, VGG16 with U-net. The trained network has been applied to the real images taken at one port in Republic of Korea, and found that the fenders are segmented with high accuracy even with a small dataset.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.05a
• /
• pp.463-466
• /
• 2010

In this paper, CCTV based fire surveillance system should aim to development. Advantages and Disadvantages analyzed of Existing sensor-based fire surveillance system and video-based fire surveillance system. To national support U-City, U-Home, U-Campus, etc, spread the ubiquitous environment appropriate to fire surveillance system model and a fire judgement technology. For this study, Microsoft LifeCam VX-1000 using through the capturing images and analyzed for apple and tomato, Finally we used H.264. The client uses the Linux OS with ARM9 S3C2440 board was manufactured, the client's role is passed to the server to processed capturing image. Client and the server is basically a 1:1 video communications. So to multiple receive to video multicast support will be a specification. Is fire surveillance system designed for multiple video communication. Video data from the RGB format to YUV format and transfer and fire detection for Y value. Y value is know movement data. The red color of the fire is determined to detect and calculate the value of Y at the fire continues to detect the movement of flame.

• Journal of The Korean Association For Science Education
• /
• v.19 no.3
• /
• pp.428-440
• /
• 1999

The colorful overhead projector films, named as Bio-cosmos II, including photographs, pictures, concept maps, and diagrams, were developed and manufactured as audio-visual teaching aids and teaching resources for students' biology learning in high school, and the CD-ROM and web sites for their application to the school were also constructed. The content of the films was organized based upon the analysis of seven different biology textbooks approved by the Ministry of Education. The films were designated based on various instructional strategies and manufactured using multimedia with various educational softwares. The CD-ROM was composed of the scenes as logo, initial main, chapters list, contents, and quit. Initial main scene indicated various chapters according to the texts of biology areas in General Science, Biology I, and II. Each chapters linked with the scenes for detailed concept maps, the downstream real subjects, and contents. The subject screens were composed of various types of summarized diagrams including lesson contents, figures, pictures, photographs, and their explanation, experimental procedures and results, tables for summarized contents, and additional animation with video captures, explanations, glossary, etc. Most files were manufactured in software Adobe Photoshop by scanning the pictures, figures and photographs, and then the explanation, modification, storing with PICT or PSD files, and transformation with JPG files, were processed in the aspect of high quality in terms of instructional strategies and graphic skills on gracefulness, clearness, colorfulness, brightness, and distinctness. A 14 films for biology areas in General Science, 80 for Biology I, and 142 for Biology II were manufactured and loaded to the CD-ROM and web site, and the files had been attempted to opened with an internet home-page of http://gic.kyungpook.ac.kr/biocosmos.