(Back to Session Schedule)

The 23rd International Technical Conference on Circuits/Systems, Computers and Communications

Session A3  Image Processing & Video Technology 3
Time: 15:30 - 17:00 Monday, July 7, 2008
Location: 10F International Conference Room
Chairs: Masahiro Okuda (The University of Kitakyushu, Japan), Yo-Sung Ho (GIST, Republic of Korea)

A3-1 (Time: 15:30 - 15:48)
TitleMonocular Vision based Indoor Mobile Robot
Author*Naoya Tada, Keisuke Murata, Takeshi Saitoh, Tomoyuki Osaki, Ryosuke Konishi (Tottori University, Japan)
Pagepp. 41 - 44
Keywordmonocular vision, indoor mobile robot, boundary detection, obstacle detection
AbstractThis paper presents the indoor mobile robot that moves automatically without needing environment information beforehand while recognizing the frontal surrounding environment with only one general camera. Based on the frontal image, the robot detects two boundary lines, some obstacle regions, and a moving direction. When the obstacle is detected, the avoidance or stop movement is worked according to the size and the position of the obstacle, otherwise the robot moves at the center of the corridor. We developed two wheelchair based mobile robots, and carried out moving experiments. It was possible to pass each other without colliding by working the avoidance movement in face-to-face movement in the coexistence environment.

A3-2 (Time: 15:48 - 16:06)
TitleSemantic-Based Scene Retrieval Using Ontologies for Video Server
Author*Min Young Jung, Sung Han Park (Hanyang University, Republic of Korea)
Pagepp. 45 - 48
Keywordvideo representation, semantic search, ontologies, scene search
AbstractIn this paper, video ontology system for retrieving a video data in a scene unit is proposed. The proposed system creates a semantic scene as a basic unit of video retrieval, and limits a domain of retrieval through a subject of that scene. Each key frames of shots contained in the semantic scene include object and event. The content of semantic scene is defined using the relationship between object and event. The semantic gap between the low level feature and the high level feature is solved through the scene ontology to ensure the semantic-based retrieval.

A3-3 (Time: 16:06 - 16:24)
TitleA Scrambling Method for JPEG Coded Images Enabling Image Retrieval from Scrambled Images
Author*Ryo Hosaka (Tokyo Metropolitan University, Japan), Fitri Arnia (Universitas Syiah Kuara, Indonesia), Masaaki Fujiyoshi, Hitoshi Kiya (Tokyo Metropolitan University, Japan)
Pagepp. 49 - 52
KeywordImage compression, Similarity measurement, Pseudo random numbers, Privacy
AbstractThis paper proposes a scrambling method for JPEG coded images and an image retrieval method for scrambled JPEG images. The proposed method scrambles JPEG images and retrieves a identical image or similar images from scrambled images. By utilizing the positive and negative signs of discrete cosine transformed coefficients as the characteristics, the method is able to retrieve the images which are compressed with the same and different ratios as the query image, without descrambling and JPEG decording.

A3-4 (Time: 16:24 - 16:42)
TitleFast Reprojection using Tree-Structured Filter Bank from Cone Beam Projections
Author*Hiroaki Morimoto, Kazuhiro Ueda, Yoshitaka Morikawa (Okayama University, Japan)
Pagepp. 53 - 56
Keyword3D X-ray CT, Cone Beam Projections, Tree-Structured Filter Bank, Reprojection
AbstractRecently, 3D X-ray CT which reconstructs 3D image receives much attention. The strict method of 3D image reconstruction from cone beam projections was proposed by Grangeat. However it requires much computation time to reconstruct 3D distribution. In this paper, the authors proposes a method using Tree-Structured Filter Bank (TSFB) to reduce processing time of reprojection which requires many calculations in Grangeat's method.

A3-5 (Time: 16:42 - 17:00)
TitleA Location Map-Free Reversible Data Hiding Method Inserting Data to Image Edges
Author*Michihiko Ono, Masaaki Fujiyoshi, Hitoshi Kiya (Tokyo Metropolitan University, Japan)
Pagepp. 57 - 60
KeywordWatermarking, Lossless, Standard deviation, Sobel filter
AbstractThis paper proposes a reversible data hiding method that embeds data to image edges. The proposed method once distorts an original image to embed data, it restores the original image as well as extracts hidden data from the image conveying hidden data. To extract data and to recover the original image, this method memorizes only one parameter and does not require any location map that records the positions conveying hidden data. Differing from the conventional method having these features, the proposed method is able to hide data to only image edges. Simulation results show the effectiveness of the proposed method.