3D Reconstruction on ShapeNet benchmark leaderboard · SOTA2 Research

3D Reconstruction on ShapeNet

85.8mIoU (Car)2D-3D self-consistency framework

Evaluation Results

Method
2D-3D self-consistency framework2018.11	85.8	—	69.2	—	—	—	56.7	—	—	—	—	—	70.6	—
Matryoshka2018.11	85	—	64.7	—	—	—	54.7	—	—	—	—	—	68.1	—
R2N22017.05	83.38	96.31	54.25	77.47	67.84	85.82	51.74	72.66	55.89	77.54	49.5	69.82	62.1	81.23
PSGN2018.11	83.1	—	60.1	—	—	—	54.4	—	—	—	—	—	65.9	—
voxel tube2018.11	82.1	—	67.1	—	—	—	55	—	—	—	—	—	68.1	—
OGN2018.11	81.6	—	58.7	—	—	—	48.3	—	—	—	—	—	62.9	—
3D-R2N22018.11	79.8	—	51.3	—	—	—	46.6	—	—	—	—	—	59.2	—
RP+NN2017.05	67.44	65.08	46.71	41.87	54.67	50.79	34.49	28.29	30.81	25.01	31.16	24.77	44.65	39.85
PTN2017.05	65.93	85.04	44.22	67.21	51.88	71.8	37.36	50.81	35.56	53.67	33.74	47.25	45.72	64.09
RP2017.05	65.21	87.13	43.44	66.94	52.42	70.23	37.17	50.48	31.97	44.64	32.1	43.77	44.42	61.23
McRecon2017.05	61.42	86.74	45.23	68.77	54.58	74.73	43.65	62.12	42.04	57.41	40.09	57.7	48.49	68.51
VC2017.05	57.84	54.3	34.52	29.36	52.57	49.41	40.48	35.09	35.49	30.11	33.87	27.88	43.36	38.57
McRecon2017.05	56.22	82.44	37.27	59.11	37.91	55.97	35.03	48.28	35.32	45.82	29.53	39.12	40.36	57.29
RP+NN2017.05	54.51	55.82	20.57	15.6	27.67	22.85	15.56	10.56	12.85	8.72	17.58	11.83	25.97	22.67
PTN2017.05	44.37	77.25	33.52	55.68	33.09	49.47	22.41	31.78	19.77	28	21.45	28.84	29.31	46.2
RP2017.05	37.91	72.5	25.08	49.97	34.27	50.93	19.3	33.61	18.21	26.64	21.88	30.03	25.77	44.52
VC2017.05	26.05	24.02	10.92	8.06	26.27	24.51	20.35	18.52	17.35	15.46	13.03	10.64	19.86	17.81