Research Article

CNN-LSTM Learning Approach-Based Complexity Reduction for High-Efficiency Video Coding Standard

Table 3

Comparative performance between the proposed approach and the state-of-the-art works.

Sequence[30][20][29]Proposed approach
BD rate (%)BD-PSNR (dB)ΔT (%)BD rate (%)BD-PSNR (dB)ΔT (%)BD rate (%)BD-PSNR (dB)ΔT (%)BD rate (%)BD-PSNR (dB)ΔT (%)

PeopleOnStreet1.05−0.045−47.505.45−0.25029.841.85−0.08527.221.70−0.017−48.88
Traffic1.99−0.052−60.603.20−0.10259.661.53−0.059−66.38
Kimono1.49−0.048−56.030.35−0.01034.742.43−0.07950.491.65−0.052−47.77
ParkScene1.47−0.042−58.722.84−0.09046.422.79−0.081−70.82
Cactus2.07−0.043−56.872.79−0.06543.222.46−0.06053.061.73−0.033−53.85
BQTerrace1.09−0.017−60.012.15−0.03838.701.74−0.03251.891.75−0.030−65.62
BasketballDrive2.268−0.052−55.842.06−0.04639.451.93−0.04649.162.02−0.045−52.77
BasketballDrill1.953−0.072−55.193.90−0.14832.122.24−0.08946.551.67−0.061−48.23
BQMall1.914−0.071−50.745.56−0.22737.111.86−0.07143.321.38−0.090−48.07
PartyScene1.011−0.039−46.834.74−0.21031.582.58−0.10830.330.96−0.038−59.30
RaceHorses0.872−0.032−46.222.10−0.18026.031.25−0.04827.081.47−0.055−54.32
BasketballPass1.45−0.066−49.043.12−0.14738.241.26−0.056−56.67
BQSquare0.77−0.028−46.913.38−0.14535.723.02−0.11734.301.27−0.046−60.79
BlowingBubbles1.29−0.044−45.623.41−0.13624.733.99−0.15439.870.97−0.034−50.14
RaceHorses1.11−0.047−41.861.60−0.018−50.33
FourPeople1.83−0.052−64.371.66−0.05865.281.83−0.06378.982.71−0.071−72.42
Johnny1.69−0.038−66.490.90−0.02064.055.45−0.13979.312.46−0.083−73.59
KristenAndSara1.55−0.045−67.231.58−0.05064.673.30−0.10877.582.93−0.094−74.91
Average1.490.04654.22.560.09943.332.970.10754.571.780.05358.60
FoM2.745.905.443.03