| Title | Mapping Functions between Image Features and KANSEI and its Application to KANSEI Based Clothing Fabric Image Retrieval |
| Author | *Shota Sobue (Ritsumeikan University, Japan), Xinyin Huang (Soochow University, China), Yen-Wei Chen (Ritsumeikan University, Japan) |
| Page | pp. 705 - 708 |
| Keyword | image feature, impression, neural network |
| Abstract | In this paper, we propose a new technique for clothing fabric image retrieval based on KANSEI (impressions). We first learn the mapping function between the fabric image features and the KANSEI and then the images in the database are projected into the KANSEI space (psychological space). The retrival is done in the psychological space by comparing the querry impression with the projection of the images in database. |
| Title | Fast Object Detection Method for Visual Surveillance |
| Author | Hyo-Kak Kim, Suryanto , *Dae-Hwan Kim, Dongni Zhang, Sung-Jea Ko (Korea University, Republic of Korea) |
| Page | pp. 709 - 712 |
| Keyword | object detection, surveillance, background-subtraction |
| Abstract | Most of the algorithms developed for object
detection employ a background-subtraction technique
which requires heavy computation. In this paper, we
present a fast background-subtraction technique which can
be readily applied to many existing object detection
algorithms. The proposed technique consists of three parts:
persistent background-subtraction, background-subtraction
with nearby searching, and skipped background-subtraction.
Experimental results show that the proposed technique can
detect the moving objects effectively without any
degradation of reliability. |
| Title | Improving the Robustness of Lips Sensing with Evolutionary Video Processing |
| Author | *Takuya Akashi, Yuji Wakasa, Kanya Tanaka (Yamaguchi University, Japan), Minoru Fukumi (The University of Tokushima, Japan) |
| Page | pp. 713 - 716 |
| Keyword | Genetic algorithm, Video processing, Human interface |
| Abstract | In this paper, an effective method is proposed for robust lips sensing. Our objectives are high-speed lips tracking and data acquisition of a talking person in natural scenes. Our approach is based on the Evolutionary Video Processing. This method has a trade-off between accuracy and a processing time. To solve this problem, we proposed automatic Search Domain Control method and implement this method in the Evolutionary Video Processing. The tracking accuracy is improved from 66.3% to 84.9%. The proposed method can recover from occlusion and tracking loss. Comparative experiments are presented to demonstrate the effectiveness and robustness of the proposed method. |
| Title | A Hybrid Technique for De-interlacing Based on Motion Compensation Reliability |
| Author | *Joonyoung Chang, Moon Gi Kang (Institute of TMS Information Technology, Yonsei University, Republic of Korea) |
| Page | pp. 721 - 724 |
| Keyword | De-interlacing, Motion compensated interpolation, Intra-field interpolation |
| Abstract | Although motion compensated interpolation (MCI) improves the
vertical resolution of de-interlaced frames effectively, it often
introduces serious defects like feathering artifacts. In this paper,
we propose an arbitration rule between MCI and intra-field
interpolation for de-interlacing to avoid the motion compensation
artifacts (MCAs) produced by erroneous MCI. In the proposed
de-interlacing method, we check the MCI results by using the
proposed MCA detection method and decide whether the MCI results are
reliable or not. And then, we use the reliability of MCI results to
combine two de-interlacing methods. The proposed arbitration method
is more elaborate than conventional methods since we directly detect
the artifacts in MCI results and the reliability of MCI results is
used as one of the important factors of the proposed arbitration
weights. Experimental results show that the proposed method achieves
better image quality than the conventional methods in terms of both
subjective and objective measures. |