TensorFlow-based implementation of "Two-phase Scheme for Trimming QTMT CU Partition using Multi-branch Convolutional Neural Networks"

Architecture

Fast QTMT Algorithm

Result

0.71% BD-BR increased / 42.341% average time saving with VTM6.1 in All-Intra configuration

Experiment

python 3.6
TensorFlow 2.0

Data

CPIH-Intra database proposed by “A Deep Convolutional Neural Network Approach For Complexity Reduction On Intra-Mode Hevc” is adopted.

Testing

Experiments are testing on all the test sequences from class A1 to class E.