Light Field Image Compression Based on Multi-view Pesudo Sequence

In recent years, Light Field Image (LFI) that can provide more immersive multimedia experience has attracted extensive attentions. Considering that the LFI has huge amount of data, this paper proposes an efficient light field image compression scheme based on multi-view pesudo sequence. In the encoder side, the raw image captured by the light field camera is firstly decomposed into sub-aperture images (SAIs) according to the lenslet array of the camera. Since the SAIs have strong intra-view and inter-view correlations, some SAIs are selected to construct the multi-view pseudo sequence, which is encoded based on MV-HEVC via the proposed prediction structure. In the decoder side, the un-encoded SAIs are interpolated based on the decoded multi-view pseudo-sequence through the video frame interpolation method, thereby restoring all light field images. Experimental results have demonstrated the proposed method outperforms the state-of-the-art LFI compression scheme based on disparity guided sparse coding, with 18.5% BD-rate reduction and 1.28 dB BD-PSNR improvement on average.