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

Table 4

The ACTS of the proposed method and the existing methods.

Sequences		Proposed	CTDM [6]	FBDA [7]	FIVG [13]	ACSD [14]
		ACTS (%)	ACTS (%)	ACTS (%)	ACTS (%)	ACTS (%)
Class B 1920 × 1080	Kimono	51.34	34.33	41.82	66.59	32.32
	ParkScene	54.31	34.56	31.60	56.28	35.41
	BQTerrace	51.51	31.07	29.47	49.44	34.50
Class C 832 × 480	PartyScene	52.43	35.93	35.23	41.71	31.10
	RaceHorsesC	50.25	36.06	33.89	52.07	23.63
	BasketballDrill	52.42	33.94	28.73	53.05	33.39
Class D 416 × 240	BlowingBubbles	51.31	35.04	21.87	43.90	33.90
	RaceHorses	47.25	35.96	31.83	44.93	31.79
	BQSquare	53.15	33.98	23.00	32.34	30.73
Class E 1280 × 720	Johnny	55.33	33.02	24.44	62.55	38.85
	FourPeople	58.45	35.05	26.65	62.18	38.01
	KristenAndSara	50.12	32.95	25.32	60.82	34.84
Average		53.32	34.32	29.49	52.16	33.21