Algorithm 1 Proposed fast CU size decision method for the low complexity H.265/HEVC encoder.

Input: CTU size = 64 × 64, the maximum quadtree depth level = 4

for Depth level=0 to 3 do

Encode the current CU with the Merge/Skip mode

Encode the current CU with the inter-2N×2N mode

if $Q_{Merge}+Q_{Inter2N\times 2N}=\text{0}$ then

The predictive distortion of the remaining inter-prediction modes is obtained by Equation (4) and Equation (5)

else

The predictive distortion of the remaining inter-prediction modes is achieved by the original motion estimation

end if

if ${\sum }_i{Q_{Merge}}_i+{\sum }_i{Q_{Inter2N\times 2N}}_i=0$ then

The CU size decision process is terminated

else

Encode the current CTU with the next quadtree depth level

end if

Output: The best CU quadtree depth level

end for

Process the next CTU