Powerful and Low Time Phase-Based Video Magnification Enhancing Technique

Abstract

Video processing techniques become major role with the wildly spread of smartphones and other new technologies. One of the high impact video processing is a magnification. Magnification refers to amplify a region or special bands of video frames in order to detect subtle information. The famous approach to achieve magnification is Eulerian-based. This paper proposes a developed approach for a phase-based Eulerian video magnification (PB-EVM) in order to reduce the processing time. The proposed approach reduces the processed data in PB-EVM by resizing-down the source video size with maintaining the video quality. Subsequently, the resulted video is decomposed using complex steerable pyramid. Then, the decomposed video is entered to temporal filters in order to pass only specific frequency bands, and then amplified the energy of these bands by multiplying by magnification factor. Finally, the video magnified video is reconstructed by steerable pyramid and resizing-up to come back to the original video size. The experimental results of the proposed approach show that there is a 60–65% reduction in the execution time compare to the conventional PB-EVM. Furthermore, the magnified video quality is maintained as it in the conventional PB-EVM.