Building Change Detection on LEVIR-CD+ (test) benchmark leaderboard · SOTA2 Research

Building Change Detection on LEVIR-CD+ (test)

91.39F1 ScoreSGSLN

Evaluation Results

Method
SGSLN2023.11	91.39	92.2	90.59	84.14
SGSLN2023.11	90.9	91.3	90.5	83.32
SGSLN2023.11	89.95	90.74	89.18	81.74
ChangeMamba (Real)2024.06	87.25	—	—	77.39
Intelligent-BCD2023.11	86.29	93.8	79.9	75.89
AttU-Net2023.11	85.99	93.5	79.6	75.43
U-Net2023.11	85.86	93.2	79.6	75.23
DinoMamba (Real)2024.06	85.14	—	—	74.12
BiT2023.11	82.79	82.74	82.85	70.64
ChangeFormer (Real)2024.06	80.58	—	—	67.48
STANet2023.11	79.27	74.62	84.54	65.66
DTCDSCN (Real)2024.06	77.63	—	—	63.44
DTCDSCN2023.11	77.6	80.36	75.03	63.4
UNet++_MSOF2023.11	75.34	85.9	67.1	60.44
FC-Siam-Diff2023.11	73.48	74.97	72.04	58.07
FC-Siam-Conc2023.11	72.97	66.24	81.22	57.44
FC-EF2023.11	66.48	61.3	72.61	49.79
DinoMamba2024.06	66.23	—	—	49.63
DinoMamba2024.06	65.46	—	—	48.78
DinoMamba2024.06	47.31	—	—	30.85
ChangeMamba trained on SynRS3D2024.06	46.78	—	—	30.39
ChangeMamba2024.06	44.3	—	—	28.28
ChangeMamba2024.06	42.5	—	—	27.5
DTCDSCN trained on SynRS3D2024.06	41.3	—	—	25.82
ChangeFormer trained on SynRS3D2024.06	38.14	—	—	23.33
ChangeFormer2024.06	38.12	—	—	23.31
DTCDSCN2024.06	35.28	—	—	21.16
ChangeFormer2024.06	27.58	—	—	15.67
DTCDSCN2024.06	21.53	—	—	11.7