Error surface-based sub-pixel accurate correction method for decoding end motion vector correction

• Main Authors: KOTCHAR, SERGEJ, THATRASRAMAN, SRIRAM, AH, CHIFFUA, RYAN
• Format: Patent
• Language: Chinese; English
• Published: 13.09.2022
• Subjects: ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; PICTORIAL COMMUNICATION, e.g. TELEVISION

In view of decoding end motion vector correction/derivation being a standard aspect of a decoding system, an encoder also needs to perform the same error plane technique to avoid any drift between encoder reconstruction and decoder reconstruction. Thus, aspects of all embodiments are applicable to encoding and decoding systems. In the template matching, the correction movement occurs only in the reference, starting from a sub-pixel precision center, which is led out based on an explicitly indicated fusion cable or is implicitly derived by a cost evaluation. In bilateral matching (with or without an averaging template), the correction begins in reference lists L0 and L1 from a respective sub-pixel precision center, where the sub-pixel precision center is derived based on a fusion index of the explicit indication or implicitly derived by a cost assessment. 鉴于解码端运动矢量修正/推导是译码系统的标准方面，编码器也需要执行相同的误差面技术，以免编码器重建与解码器重建之间有任何漂移。因此，所有实施例的各方面都适用于编码系统和解码系统。在模板匹配中，修正移动仅在参考中发生，从子像素精确中心开始，其中，所述子像素精确中心是基于显式指示的融合索引导出的或通过代价评估隐式推导