The Multi-Reference Minimum Rate Predictors (M-MRP) algorithm is a lossless encoder which presents one of the highest compression efficiency for Light Field encoding, but its high computational complexity is still a concern. This article describes an experimental study on the trade-off between the encoding performance and computational complexity of M-MRP Light Field encoding, taking the HEVC Reference Software as a benchmarking reference.
The results show that even with an average complexity reduction of approximately 40 times, M-MRP still achieves better compression ratios than other state-of-the-art encoders, such as HEVC or JPEG-LS. Furthermore, for a similar encoding complexity M-MRP outperforms the HEVC reference software by 13.2% on average.