List your company
ResearchFormerly WizWand

Image Classification on PET

93.4AccuracyChimeraLoRA
49.241660.705872.1783.6342Apr 22, 2025Jun 12, 2025Aug 2, 2025Sep 22, 2025Nov 12, 2025Jan 2, 2026Feb 23, 2026
Updated 6mo ago

Evaluation results

MethodLinks
ChimeraLoRAEvaluation Protocol=Few-shot (Augmented), Shot Count=4 Real + 500 Synthetic2026.0293.4
CLIPEvaluation Protocol=Few-shot (Real-only), Shot Count=42026.0293.3
LoFTEvaluation Protocol=Few-shot (Augmented), Shot Count=4 Real + 500 Synthetic2026.0292.4
DataDreamEvaluation Protocol=Few-shot (Augmented), Shot Count=4 Real + 500 Synthetic2026.0292.2
Ours (T*cn)Backbone=SwinT, Training Scenario=FT2025.0492.15
Ours (T*)Backbone=SwinT, Training Scenario=FT2025.0491.71
Ours (T*csgcn)Backbone=SwinT, Training Scenario=FT2025.0491.63
IsSynthEvaluation Protocol=Few-shot (Augmented), Shot Count=4 Real + 500 Synthetic2026.0291.6
T = sqrt(M)Backbone=SwinT, Training Scenario=FT2025.0491.39
Insert LNBackbone=SwinT, Training Scenario=FT2025.0491.2
Ours (T*csg)Backbone=SwinT, Training Scenario=FT2025.0491.14
Ours (T*csgcn)Backbone=ResNet50, Training Scenario=FT2025.0489.81
Ours (T*)Backbone=ResNet50, Training Scenario=FT2025.0489.59
T = 1Backbone=SwinT, Training Scenario=FT2025.0489.4
Ours (T*cn)Backbone=ResNet50, Training Scenario=FT2025.0489.26
CLIPEvaluation Protocol=Zero-shot, Shot Count=02026.0289.2
Ours (T*csg)Backbone=ResNet50, Training Scenario=FT2025.0488.83
Insert LNBackbone=ResNet50, Training Scenario=FT2025.0488.09
T = 1Backbone=ResNet50, Training Scenario=FT2025.0487.44
Ours (T*csg)Backbone=ResNet50, Training Scenario=init2025.0477.49
LGCLIPBackbone=ViT-L/14, Correction (Manual Calibration)=false2025.1277.43
T = sqrt(M)Backbone=ResNet50, Training Scenario=FT2025.0477.19
Ours (T*)Backbone=ResNet50, Training Scenario=init2025.0476.12
T = sqrt(M)Backbone=ResNet50, Training Scenario=init2025.0476.02
Ours (T*csgcn)Backbone=ResNet50, Training Scenario=init2025.0475.63
Ours (T*cn)Backbone=ResNet50, Training Scenario=init2025.0474.03
CLIPBackbone=ViT-L/14, Correction (Manual Calibration)=false2025.1272.74
Insert LNBackbone=ResNet50, Training Scenario=init2025.0471.46
LGCLIPBackbone=ViT-B/16, Correction (Manual Calibration)=false2025.1267.18
CLIPBackbone=ViT-B/16, Correction (Manual Calibration)=false2025.1265.14
LGCLIPBackbone=ViT-B/32, Correction (Manual Calibration)=false2025.1263.75
CLIPBackbone=ViT-B/32, Correction (Manual Calibration)=false2025.1260.26
LGCLIPBackbone=RN50, Correction (Manual Calibration)=false2025.1257.05
T = 1Backbone=ResNet50, Training Scenario=init2025.0455.74
CLIPBackbone=RN50, Correction (Manual Calibration)=false2025.1250.94