A Hybrid Compression Method for Medical Images Based on Region of Interest Using Artificial Neural Networks

Table 3

Summary of the performance measures for the proposed compression method at ∝ = 0.8.

Category	This study ()				Ahamed et al. [16]
	ROI	NROI	Entire image
	PSNR (dB)	PSNR (dB)	CR	PSNR (dB)	BP		GS/PS
					CR	PSNR (dB)	CR	PSNR (dB)
Neck	58.68	45.77	7.35	49.61	4.92	53	6.07	46.24
Chest	59.61	45.47	5.74	49.19	5.06	52	6.12	47.04
Front skull view	55.01	46.11	6.94	50.77	5.27	54	7.89	41.99
Right skull view	53.78	45.44	5.84	49.44	5.29	53	7.94	40.87
Left skull view	54.05	44.68	6.70	47.88	5.76	54	7.97	41.91
Pelvic girdle	58.43	46.72	6.92	52.00	5.13	52	8.28	43.41
Radius and ulna bones	58.48	44.93	6.67	48.72	4.95	55	8.44	42.76
Pelvic girdle and back bone	55.36	43.94	6.22	48.10	4.87	52	8.89	40.57
Hand x-ray	55.12	43.99	7.17	47.21	4.85	51	9.02	39.83
Leg	56.79	44.22	6.90	48.46	4.85	52	9.04	38.93
Left mammogram	54.21	45.08	6.84	48.62	6.02	56	13.9	37.09
Right mammogram	53.31	45.11	7.49	48.80	5.95	56	13.8	36.84
Average	54.27	45.12	6.73	49.06	5.2	53.33	8.85	41.46