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

Table 2

The results of the proposed method.

Sequences		Proposed method
(%)		BDBR	ACTS
Class A; 2560 × 1600	PeopleOnStreet	0.96	52.13
	Traffic	0.97	50.04
	Nebuta	0.99	55.27
Class B; 1920 × 1080	Kimono	1.34	51.34
	ParkScene	0.95	54.31
	BQTerrace	1.22	51.51
Class C; 832 × 480	PartyScene	0.95	52.43
	RaceHorsesC	0.96	50.25
	BasketballDrill	0.89	52.42
Class D; 416 × 240	BlowingBubbles	0.95	51.31
	RaceHorses	0.86	47.25
	BQSquare	0.87	53.15
Class E; 1280 × 720	Johnny	0.88	55.33
	FourPeople	0.78	58.45
	KristenAndSara	1.21	50.12
	Average	0.99	52.35