Benchmark Hair Removal (GAN)

Hair Removal benchmark using hairless samples from the PH2 database and the GAN-based deep neural network by Attia et al. to simulate hair on them.

Method

Hair detection and removal for dermoscopic images.

Sample	Vifp	Uqi	Msssim	P hvs m	P hvs	Mse	Ssim	Psnr	Rmse
IMD006	0.977	0.999	0.947	35.252	34.177	28.070	0.828	33.648	5.298
IMD010	0.971	0.998	0.934	32.522	31.453	51.019	0.832	31.054	7.143
IMD017	0.973	0.997	0.961	34.815	33.562	30.787	0.849	33.247	5.549
IMD018	0.981	0.999	0.955	37.874	36.481	60.419	0.845	30.319	7.773
IMD019	0.969	0.999	0.935	30.304	29.403	59.452	0.836	30.389	7.711
IMD020	0.971	0.999	0.951	35.276	33.645	67.234	0.819	29.855	8.200
IMD030	0.974	0.999	0.950	30.900	30.028	53.062	0.831	30.883	7.284
IMD033	0.968	0.999	0.936	32.673	31.582	45.577	0.813	31.543	6.751
IMD044	0.937	0.992	0.882	28.455	27.775	101.895	0.728	28.049	10.094
IMD050	0.979	0.999	0.955	34.913	34.079	30.624	0.841	33.270	5.534
IMD061	0.954	0.996	0.929	28.315	27.389	139.485	0.798	26.686	11.810
IMD063	0.971	0.997	0.963	28.731	28.102	80.223	0.826	29.088	8.957
IMD075	0.971	0.996	0.950	27.059	26.456	103.663	0.814	27.975	10.181
Mean	0.969	0.998	0.942	32.084	31.087	65.501	0.820	30.462	7.868