List your company

Image Captioning Evaluation on Flickr8K Expert (test)

56.4Kendall tau_cPolos
-1.907613.2299528.367543.50505May 25, 2022Sep 26, 2022Jan 29, 2023Jun 2, 2023Oct 5, 2023Feb 6, 2024Jun 10, 2024

Evaluation results

MethodLinks
PolosType=reference+image, Explainable=false2024.0656.4——
RefPAC-SType=Reference-based2023.0355.9—55.5
RefPAC-SType=reference+image, Explainable=false2024.0655.9——
InfoMetIC+Type=reference-free, Explainable=false2024.0655.5——
CLIP + DN-refSetting=Reference-based2023.0255.3——
MID2022.0554.9——
MIDType=Hybrid2023.0354.9——
PAC-SType=Reference-free2023.0354.3—53.9
CLIP + DNSetting=Reference-free2023.0254.3——
CLIP + DN*-refSetting=Reference-based2023.0254.3——
PAC-SType=reference-free, Explainable=false2024.0654.3——
CLIP + TTA + DNSetting=Reference-free2023.0253.6——
CLIP + TTA + DN*Setting=Reference-free2023.0253.5——
CLIP + DN*Setting=Reference-free2023.0253.2——
RefCLIP-S2022.0553——
RefCLIP-SType=Reference-based2023.0353—52.6
CLIP-refSetting=Reference-based2023.0253——
RefCLIPScoreType=reference+image, Explainable=false2024.0653——
FLEURType=reference-free, Explainable=true2024.0653——
CLIP + TTASetting=Reference-free2023.0251.9——
RefFLEURType=reference+image, Explainable=true2024.0651.9——
CLIPSetting=Reference-free2023.0251.4——
CLIP-S2022.0551.2——
CLIP-SType=Reference-free2023.0351.2—51.1
CLIPScoreType=reference-free, Explainable=false2024.0651.2——
ViLBERTScore-F2022.0550.1——
ViLBERTScoreType=Reference-based2023.0350.1——
ViLBERTScore-FSetting=Reference-based2023.0250.1——
VILBERTScore-FType=reference+image, Explainable=false2024.0650.1——
NUBIA2022.0549.5——
NUBIA*Setting=Reference-based2023.0249.5——
TIGEr2022.0549.3——
TIGErType=Reference-based2023.0349.3——
TIGErType=reference+image, Explainable=false2024.0649.3——
CLAIRType=reference-only, Explainable=true2024.0648.3——
UMICType=Reference-based2023.0346.8——
BERT-S++2022.0546.7——
BERT-S++Type=Reference-based2023.0346.7——
LEICcoefficient_type=tau_b2022.0546.6——
SPICE2022.0544.9——
SPICEType=Reference-based2023.0344.9—51.7
SPICEType=reference-only, Explainable=false2024.0644.9——
CIDEr2022.0543.9——
CIDErType=Reference-based2023.0343.9—43.6
CIDErSetting=Reference-based2023.0243.9——
CIDErType=reference-only, Explainable=false2024.0643.9——
TCL + DNSetting=Reference-free2023.0242——
METEOR2022.0541.8——
METEORType=Reference-based2023.0341.8—41.5
METEORType=reference-only, Explainable=false2024.0641.8——
BERT-Sbackbone=RoBERTa-F2022.0539.2——
BERT-SType=Reference-based2023.0339.2——
BERTScoreType=reference-only, Explainable=false2024.0639.2——
TCL + DN*Setting=Reference-free2023.0236——
ALBEF + DNSetting=Reference-free2023.0234.8——
BLEU-12022.0532.3——
ROUGE-L2022.0532.3——
BLEU-1Type=Reference-based2023.0332.3—32.2
ROUGEType=Reference-based2023.0332.3—32.1
BLEU-1Setting=Reference-based2023.0232.3——
ROUGE-LType=reference-only, Explainable=false2024.0632.3——
TCLSetting=Reference-free2023.0231——
BLEU-42022.0530.8——
BLEU-4Type=Reference-based2023.0330.8—30.6
BLEU-4Setting=Reference-based2023.0230.8——
BLEU-4Type=reference-only, Explainable=false2024.0630.8——
ALBEF + DN*Setting=Reference-free2023.0229.2——
ALBEFSetting=Reference-free2023.0224.9——
FACTUAL-T5(pre)Metric=SoftSPICE2023.050.53350.6952—
SPICE-ParserMetric=SoftSPICE2023.050.51890.6811—
CDP-T5Metric=SoftSPICE2023.050.45540.5964—
FACTUAL-T5(pre)Metric=SPICE2023.050.45120.6078—
SPICE-ParserMetric=SPICE2023.050.44770.6011—
VG-T5Metric=SoftSPICE2023.050.39660.5305—
VG-T5Metric=SPICE2023.050.37180.5194—
CDP-T5Metric=SPICE2023.050.3350.495—
LEICType=Reference-based2023.03——46.6