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

Second

Benchmarks

Task NameDataset NameSOTA ResultTrend
Semantic Change DetectionSECOND
mIoU74.33
48
Change DetectionSECOND (test)
F1 Score48.2
20
Building Change DetectionSECOND
IoU67.65
16
Object Change ProposalSECOND binary
F1 Score69.5
14
Open-Ended Visual Question AnsweringDelta-SECOND
H-CQS METEOR32.74
13
Choice Visual Question AnsweringDelta-SECOND
H-CIC Single Accuracy64.18
13
Semantic Change DetectionSECOND (test)
Params (M)2.97
13
Building Change DetectionSECOND Building 1.0 (test)
Precision79.9
12
Speech EnhancementSecond (test)
CSIG4.38
11
Reference-based Super-ResolutionSECOND 16x (test)
LPIPS0.3433
8
Reference-based Super-ResolutionSECOND 8x (test)
LPIPS0.2642
8
Reference-based Super-ResolutionSECOND full x16 (test)
LPIPS0.2905
7
Reference-based Super-ResolutionSECOND full x8 (test)
LPIPS0.2174
7
Semantic-mask-conditioned Change GenerationSECOND
mIoU76.19
6
Open-vocabulary change detectionSECOND (test)
Building IoU48.91
5
Land Cover Change DetectionSECOND (test)
Building IoU45.2
5
Building Change DetectionSECOND (test)
Precision79.9
5
Showing 17 of 17 rows