Share your thoughts, 1 month free Claude Pro on usSee more

Semantic Segmentation on NYU-depth v2 (IoU and Pixel Accuracy)

58.8mIoU

RSGMamba-B

Updated 4d ago

Evaluation Results

Method	Links
RSGMamba-B 2026.04		58.8	-	-
DFormerV2 2026.04		58.4	-	-
DFNet 2026.04		57.9	-	-
PrimKD 2026.04		57.8	-	-
ADBNet 2026.04		57.6	-	-
Sigma 2026.04		57	-	-
CMX 2026.04		56.9	-	-
CMNext 2026.04		56.9	-	-
GeminiFusion 2026.04		56.8	-	-
RSGMamba-S 2026.04		56.4	-	-
CMX 2026.04		56.3	-	-
DiffPixelFormer 2026.04		56.3	-	-
DFNet 2026.04		56.1	-	-
ADBNet 2026.04		56	-	-
ECMRN 2026.04		54.6	-	-
TokenFusion 2026.04		54.2	-	-
Sigma 2026.04		53.9	-	-
DCANet 2026.04		53.3	-	-
RefineNet-LW-101 2021.07		43.6	-	-
TD2-PSP50 2021.07		43.5	-	-
CI-Net 2021.07		42.6	-	72.7
3DGNN 2021.07		42	-	-
D-CNN 2021.07		41	-	-
Context 2021.07		40.6	-	70
Eigen et al. 2021.07		34.1	-	65.6
B-SegNet 2021.07		32.4	-	68
Deng et al. 2021.07		31.5	-	63.8
FCN 2021.07		29.2	-	60
baseline 2017.05		-	40.6	70.3
w/ gt-depth 2017.05		-	41.3	70.8
w/ pred-depth 2017.05		-	41.8	71.1
loop1 w/o depth 2017.05		-	41.9	70.6
loop1 w/ gt-depth 2017.05		-	42.5	71.1
loop1 w/ pred-depth 2017.05		-	42.7	71.2
loop2 2017.05		-	43.1	71.3
loop2 (test-aug) 2017.05		-	44.5	72.1
Context 2017.05		-	40.6	70
RefineNet-Res50 2017.05		-	43.8	-
RefineNet-Res101 2017.05		-	44.7	-
RefineNet-Res152 2017.05		-	46.5	73.6
Single-task baseline 2018.01		-	33.5	64.1
Multiple-task baseline 2018.01		-	33.4	64.2
Cross-stitch network 2018.01		-	34.8	65
Sluice network 2018.01		-	34.9	65.2
NDDR-CNN 2018.01		-	36.2	66.4
Single-task baseline 2018.01		-	33.5	64.4
Multi-task baseline 2018.01		-	33.6	64.4
Cross-stitch network 2018.01		-	35	65.1
Sluice network 2018.01		-	35.3	65.3
NDDR-CNN-Shortcut 2018.01		-	36.7	67