| Title |
Foreground-aware Padding-based Video Retargeting |
| Authors |
김정현(Jeong Hyeon Kim) ; 조성인(Sung In Cho) |
| DOI |
https://doi.org/10.5573/ieie.2019.56.5.29 |
| Keywords |
Video retargeting ; Foreground detection ; Logo detection ; Padding |
| Abstract |
In this paper, we propose a new video retargeting method that can preserve visually important areas while perfectly maintain the temporal coherence by using the foreground-aware padding with a region-adaptive scaling technique. Specifically, in the proposed method, the padding region for a given image is determined considering the results of the logo and foreground detections. After the foreground-aware padding, the region adaptive scaling is applied on a given image depending on the foreground detection results to minimize spatial distortion of foreground regions. Experimental results showed that the proposed method successfully reduced spatial distortions of visually important content compared to conventional methods and maintained temporal coherence perfectly. For the objective evaluation, the proposed method improved the bidirectional similarity value, which can measure the quality of video retargeting, by up to 1.57 compared with the benchmark methods. |