A Novel Approach for Video Frame Interpolation using Cubic Motion Compensation Technique

Increasing frame rates by frame interpolation is one of the main challenges in video processing. Motion estimation and motion compensation are two main keys of video frame interpolation algorithm. Motion estimation algorithms are used to get refine motion vectors while motion compensation algorithm assigns motion vectors accurately. Recent studies on video frame interpolation mainly focused on motion estimation algorithms, but along with that assigning motion vectors accurately is also an important issue. This paper concentrates on motion compensation. In this paper we have proposed a novel method for video frame interpolation using Cubic Motion Compensation technique which assigns motion vectors accurately. Instead of using two consecutive frames for motion estimation as in Conventional method, this paper considers four consecutive frames and calculates three different motion vectors. It is observed that proposed algorithm have relatively outperforms over conventional Motion compensated frame interpolation algorithm. For evaluation, average Peak Signal to Noise Ratio (PSNR) and average Structural Similarity (SSIM) are considered to compare the results. According to simulation results, using test sequences of different frame sizes, it can be observed that the proposed algorithm have effectively improves average PSNR of reconstructed frame by around 3dB and average SSIM have increased upto 5% with reference to the traditional methods.