Online learning early skip decision method for the HEVC Inter process using the SVM-based Pegasos algorithm

High efficient video coding (HEVC), the latest coding standard, has an encoding complexity much higher compared with H.264/advanced video coding (AVC). The greater efficiency in HEVC is obtained at much greater computational cost compared with AVC. A fast coding unit (CU) splitting algorithm is proposed for the HEVC encoder, which early terminates the CU partitioning process based on an adaptive classification model. This model is generated by an online learning method based on a Pegasos (primal estimated sub-gradient solver for SVM) algorithm. The proposed method is implemented over the HEVC reference implementation on its version 16.7. Experimental results show that the proposed method reduces the computational complexity of HEVC encoder to 35% without any loss, resulting in a 1% of Bjontegaard Delta-rate gain in the low delay B configuration without any offline training phase.