Research Article

Fast CU Size Decision Method Based on Just Noticeable Distortion and Deep Learning

Table 3

The BDBR of the proposed method and the latest methods.

SequencesProposedCTDM [6]FBDA [7]FIVG [13]ACSD [14]
BDBR (%)BDBR (%)BDBR (%)BDBR (%)BDBR (%)

Class B
1920 × 1080
Kimono1.341.051.981.720.87
ParkScene0.951.111.381.280.83
BQTerrace1.221.001.191.160.95

Class C
832 × 480
PartyScene0.950.761.050.280.55
RaceHorsesC0.960.822.960.840.37
BasketballDrill0.891.671.361.911.30

Class D
416 × 240
BlowingBubbles0.950.740.730.490.95
RaceHorses0.860.951.590.540.71
BQSquare0.870.61−0.110.170.68

Class E
1280 × 720
Johnny0.881.441.513.071.72
FourPeople0.781.381.372.551.35
KristenAndSara1.211.191.532.561.61

Average0.991.061.381.380.99