GAN Evaluation Report - Patient PD04012016

Native→Venous Translation Metrics (Generated Venous vs Real Venous)

Targeted Slice 70 - Network-Normalized Analysis (Generated vs Real Venous)

0.781

SSIM Score

0.163

RMSE

0.075

MAE

Average Network-Normalized Metrics Across All Slices (29 slices) - Generated vs Real Venous

0.807

SSIM Score (Avg)

0.138

RMSE (Avg)

0.064

MAE (Avg)

Targeted Slice 70 - HU-Space Analysis (Generated vs Real Venous)

0.781

HU SSIM

162.9

HU RMSE

75.1

HU MAE

Average HU-Space Metrics Across All Slices (29 slices) - Generated vs Real Venous

0.807

HU SSIM (Avg)

137.8

HU RMSE (Avg)

63.6

HU MAE (Avg)

Targeted Slice 70 - Lung Window Analysis (Generated vs Real Venous)

0.672

Lung SSIM

151.5

Lung RMSE

64.6

Lung MAE

Average Lung Window Metrics Across All Slices (29 slices) - Generated vs Real Venous

0.693

Lung SSIM (Avg)

121.2

Lung RMSE (Avg)

51.6

Lung MAE (Avg)

Targeted Slice 70 - Mediastinum Window Analysis (Generated vs Real Venous)

0.685

Mediastinum SSIM

54.2

Mediastinum RMSE

23.6

Mediastinum MAE

Average Mediastinum Window Metrics Across All Slices (29 slices) - Generated vs Real Venous

0.699

Mediastinum SSIM (Avg)

48.9

Mediastinum RMSE (Avg)

20.4

Mediastinum MAE (Avg)

Detailed Per-Slice Metrics

Slice ID	SSIM	RMSE	MAE	HU SSIM	HU RMSE	HU MAE	Lung SSIM	Lung RMSE	Lung MAE	Mediastinum SSIM	Mediastinum RMSE	Mediastinum MAE
PD04012016_slice_1	0.804	0.160	0.069	0.804	160.2	69.1	0.697	140.8	53.9	0.690	47.4	19.8
PD04012016_slice_2	0.809	0.160	0.069	0.809	160.3	68.6	0.698	139.0	53.0	0.690	47.5	19.3
PD04012016_slice_3	0.807	0.159	0.068	0.807	159.3	68.2	0.693	136.9	52.9	0.687	48.9	20.1
PD04012016_slice_5	0.813	0.156	0.067	0.813	156.3	66.9	0.702	132.3	51.0	0.691	49.0	20.1
PD04012016_slice_9	0.815	0.147	0.064	0.815	147.0	63.5	0.710	123.7	47.9	0.698	48.9	19.6
PD04012016_slice_10	0.815	0.146	0.064	0.815	146.1	63.7	0.708	124.5	48.9	0.699	49.5	19.8
PD04012016_slice_13	0.809	0.133	0.061	0.809	133.0	60.9	0.694	114.6	48.5	0.698	49.2	20.5
PD04012016_slice_15	0.812	0.127	0.059	0.812	127.0	59.3	0.698	108.0	46.8	0.703	49.3	20.5
PD04012016_slice_18	0.831	0.121	0.056	0.831	120.9	56.0	0.718	101.6	43.6	0.721	45.0	17.7
PD04012016_slice_22	0.830	0.111	0.054	0.830	110.8	54.4	0.708	95.9	44.3	0.728	44.2	18.1
PD04012016_slice_23	0.824	0.112	0.056	0.824	111.8	55.6	0.702	97.1	45.5	0.726	45.1	18.6
PD04012016_slice_25	0.828	0.114	0.056	0.828	113.8	55.8	0.710	96.3	44.5	0.729	44.5	17.7
PD04012016_slice_34	0.816	0.117	0.058	0.816	116.8	57.6	0.695	100.3	46.7	0.723	45.2	18.4
PD04012016_slice_35	0.818	0.120	0.058	0.818	119.6	58.0	0.698	101.7	46.5	0.723	44.7	18.0
PD04012016_slice_36	0.815	0.121	0.059	0.815	120.5	58.6	0.696	102.5	46.9	0.722	44.6	18.0
PD04012016_slice_37	0.813	0.121	0.059	0.813	121.1	58.7	0.693	103.7	47.1	0.717	45.4	18.3
PD04012016_slice_41	0.782	0.123	0.062	0.782	123.2	62.0	0.659	108.7	51.3	0.699	49.7	21.6
PD04012016_slice_51	0.807	0.126	0.061	0.807	126.5	61.4	0.689	112.1	50.9	0.703	47.8	19.9
PD04012016_slice_52	0.802	0.128	0.062	0.802	127.5	62.2	0.685	113.0	51.6	0.699	48.7	20.5
PD04012016_slice_55	0.815	0.130	0.061	0.815	130.2	61.2	0.702	113.9	49.8	0.706	46.6	18.6
PD04012016_slice_58	0.810	0.130	0.061	0.810	129.8	61.1	0.696	112.6	49.6	0.702	47.8	19.3
PD04012016_slice_61	0.802	0.131	0.062	0.802	131.3	62.1	0.687	115.8	51.5	0.704	48.4	20.1
PD04012016_slice_63	0.798	0.134	0.064	0.798	134.1	64.0	0.680	118.3	53.0	0.702	48.7	20.5
PD04012016_slice_65	0.786	0.147	0.069	0.786	147.4	68.7	0.673	132.7	57.4	0.691	52.6	22.5
PD04012016_slice_66	0.780	0.146	0.069	0.780	145.6	69.0	0.669	131.7	58.0	0.688	54.1	23.3
PD04012016_slice_70	0.781	0.163	0.075	0.781	162.9	75.1	0.672	151.5	64.6	0.685	54.2	23.6
PD04012016_slice_73	0.775	0.184	0.083	0.775	184.3	82.8	0.670	174.0	72.5	0.669	58.0	25.8
PD04012016_slice_84	0.804	0.170	0.071	0.804	170.2	71.3	0.698	160.6	61.5	0.647	57.5	25.2
PD04012016_slice_90	0.808	0.160	0.067	0.808	159.9	67.2	0.697	150.6	57.7	0.634	56.3	25.1

Image Grid

4×3 grid: Rows show different image types (Original NATIVE, Reconstructed NATIVE, Original VENOUS, Generated VENOUS), Columns show windowing techniques (No Window, Lung Window, Mediastinum Window)

Original NATIVE CT scan (input)

No window - Raw intensity values