A digital system for efficient de-interlacing based on refined motion adaptive method is presented. The adaptive algorithm consists of same parity four-field motion detection, temporal averaging “Weave”, and edged-based line averaging (ELA). Four motion-levels can be estimated to avoid the influence of noise, while the usage of external memory for field buffering is saved. Adaptive fading between “Weave” and ELA is adopted to reconstruct image effectively with the estimated multilevel coefficient of motion. In refined ELA module, the accuracy of edge directional interpolation is ensured by condition judgment and de-ambiguous processing, while the efficiency of algorithm is improved under four-stage pipeline management. The implemented IC achieves high computational efficiency and low operating power consumption with reasonable architecture strategies. Experimental results show both excellent subjective quality and objective performance for real-time video applications.
展开▼
