2D Object Detection on SUN-RGBD (val)

0.881AP (bed)UODDM

Evaluation Results

Method
UODDM2022.07	0.881	0.876	0.538	0.668	0.695	0.287	0.622	0.472	0.197	0.419	—	68.5	35.8	28.5	62.8	41.9	0.566	53.4	—	—
UODDM + CPPM2022.07	0.88	0.856	0.518	0.683	0.686	0.269	0.616	0.455	0.202	0.417	—	67.6	33	29.2	61.6	47.4	0.558	52.8	—	—
Ours with simCrossTrans2022.03	0.872	0.877	0.516	0.695	0.69	0.27	0.605	0.481	0.193	0.366	61.2	35.5	47.7	41.9	55.8	52.7	—	—	—	—
simCrossTrans2022.07	0.872	0.877	0.516	0.695	0.69	0.27	0.605	0.481	0.193	0.383	—	68.1	35.5	30.7	61.2	41.9	0.558	52.7	—	—
UODDM + CPPM2022.07	0.865	0.91	0.544	0.702	0.672	0.303	0.575	0.487	0.228	0.527	—	66.6	41.2	29.2	68.6	57.9	0.581	55.8	—	—
UODDM2022.07	0.836	0.871	0.533	0.588	0.625	0.226	0.542	0.468	0.22	0.48	—	63.7	38.4	28.9	67.1	57.4	0.539	52.5	—	—
UODDM + CPPM2022.07	0.836	0.886	0.53	0.591	0.608	0.265	0.507	0.461	0.22	0.52	—	60.6	37.8	31.2	64.7	54.3	0.542	51.9	—	—
rgb only model2022.07	0.832	0.939	0.518	0.542	0.604	0.237	0.513	0.463	0.225	0.544	—	60.4	39.8	32.7	67	48.1	0.542	52.3	—	—
Ours with simCrossTrans2022.03	0.821	0.788	0.435	0.499	0.6	0.148	0.503	0.411	0.13	0.316	58.9	26.7	41.4	29.8	25.6	30.6	0.393	46.5	—	—
F-VoxNet2022.03	0.81	0.895	0.351	0.5	0.524	0.219	0.531	0.377	0.183	—	—	—	—	—	—	—	0.568	—	—	—
F-VoxNet2022.07	0.81	0.895	0.351	0.5	0.524	0.219	0.531	0.377	0.183	0.404	—	47.8	29.8	22	52.8	39.7	0.479	—	—	—
Frustum VoxNet2019.10	0.81	0.895	0.351	0.5	0.524	0.219	0.531	0.377	0.183	0.404	—	22	0.528	—	39.7	0.31	—	—	47.8	29.8
F-VoxNet2022.03	0.787	0.776	0.342	0.519	0.518	0.165	0.485	0.349	0.142	—	—	—	—	—	—	—	0.419	—	—	—
F-VoxNet2022.07	0.787	0.776	0.342	0.519	0.518	0.165	0.485	0.349	0.142	0.192	—	48.7	18.5	19.1	30.3	22.2	0.428	37.3	—	—
Frustum VoxNet2019.10	0.787	0.776	0.342	0.519	0.518	0.165	0.485	0.349	0.142	0.192	—	19.1	0.303	—	22.2	0.301	—	—	48.7	18.5
Ours w/o simCrossTrans2022.03	0.767	0.584	0.231	0.497	0.572	0.121	0.405	0.415	0.135	0.197	38.3	68.1	37.8	30.7	29.2	23.7	0.252	39.6	—	—
RGB-D RCNN2022.03	0.76	0.698	0.371	0.496	0.412	0.313	0.422	0.43	0.166	0.349	—	—	—	—	—	—	0.457	—	—	—
RGB-D RCNN2022.07	0.76	0.698	0.371	0.496	0.412	0.313	0.422	0.43	0.166	0.349	—	—	—	—	46.8	41.9	0.442	—	—	—
RGB-D RCNN2019.10	0.76	0.698	0.371	0.496	0.412	0.313	0.422	0.43	0.166	0.349	—	—	0.468	—	—	0.419	—	—	—	—
2D-driven2022.03	0.745	0.862	0.495	0.455	0.53	0.294	0.49	0.423	0.223	0.672	—	—	—	—	—	—	0.468	—	—	—
2D-driven2022.07	0.745	0.862	0.495	0.455	0.53	0.294	0.49	0.423	0.223	0.457	—	—	—	—	—	—	0.497	—	—	—
2D-driven2019.10	0.745	0.862	0.495	0.455	0.53	0.294	0.49	0.423	0.223	0.457	—	—	—	—	—	—	—	—	—	—
Ours w/o simCrossTrans2022.03	0.722	0.517	0.22	0.444	0.495	0.096	0.336	0.335	0.097	0.122	41.8	15.7	25.3	22.8	22.8	12.9	0.195	33.8	—	—
Frustum PointNets2022.03	0.567	0.435	0.372	0.813	0.641	0.333	0.574	0.499	0.778	—	—	—	—	—	—	—	0.497	—	—	—
Frustum PointNets2022.07	0.567	0.435	0.372	0.813	0.641	0.333	0.574	0.499	0.778	0.672	—	—	—	—	—	—	0.568	—	—	—
Frustum PointNets2019.10	0.567	0.435	0.372	0.813	0.641	0.333	0.574	0.499	0.778	0.672	—	—	—	—	—	—	—	—	—	—