Adaptive frame rate up-conversion based on motion classification
Signal Processing
Fuzzy rough sets hybrid scheme for motion and scene complexity adaptive deinterlacing
Image and Vision Computing
IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
Multiresolution-Based Texture Adaptive Algorithm for High-Quality Deinterlacing
IEICE - Transactions on Information and Systems
De-interlacing algorithm using spatial-temporal correlation-assisted motion estimation
IEEE Transactions on Circuits and Systems for Video Technology
Global motion compensation based de-interlacing using adaptive integral projection
SIP '07 Proceedings of the Ninth IASTED International Conference on Signal and Image Processing
| Hi-index | 0.43 |
Motion estimation (ME) and motion compensation (MC) based deinterlacing has been become the mainstream technology of industry. The existing ME/MC algorithms are mostly based on block matching (BM) strategy, in which various BM algorithms are proposed with different tradeoff between the matching accuracy and the CPU time. Aiming at the dilemma between the performance and the efficiency of the available deinterlacing algorithms, two adaptive ME/MC algorithms are presented in this paper. Compared with the full search BM algorithm, the proposed algorithms greatly reduce the computational amount with keeping the performance approximately. However, the inherent drawbacks of the BM limit the farther improvement of the deinterlacing performance. For this purpose, a new deinterlacing algorithm based on motion object is developed, in which it is natural motion object rather than contrived block that is taken as the basic cell for ME/MC. In the new algorithm, a more reliable technique is adopted for accurate motion region detection, and immune clonal selection algorithm is employed to accelerate the matched object searching. Meanwhile, many advanced deinterlacing methods, such as motion compensation, median filtering, Weave and Bob, are integrated into the new algorithm, so it is more adaptive to the various video sequences. The experimental results with several available test databases illustrate that the new algorithm based on motion object is better than those based on motion block on the whole.