AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

SPICE: Semantic Propositional Image Caption Evaluation

Peter Anderson, Basura Fernando, Mark Johnson, Stephen Gould

ECCV 2016 · Vision-language and multimodal › Human and embodied modeling · approach: Neural image captioning · 2,045 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use SPICE instead of CIDEr for judging captionsSPICE · COCO 2015 Challenge system-levelPearson correlation with human M1: 0.43 -> 0.88Improved
use SPICE instead of METEOR for judging captionsSPICE · COCO 2015 Challenge system-levelPearson correlation with human M1: 0.53 -> 0.88Improved
add synonym matching to exact-match SPICE (ablation)SPICE · COCO 2015 Challenge system-levelPearson correlation with human M1: 0.84 -> 0.88Improved
use SPICE instead of CIDEr on detail judgmentsSPICE · COCO 2015 Challenge system-levelPearson correlation with human M4 (detail): 0.46 vs -0.21Improved
use SPICE instead of CIDEr at caption level on Flickr 8KSPICE · Flickr 8K caption-levelKendall tau with human: 0.44 -> 0.45Improved
use SPICE instead of CIDEr at caption level on the composite datasetSPICE · Composite dataset caption-levelKendall tau with human: 0.36 -> 0.39Improved
use SPICE instead of METEOR on pairs of model captionsSPICE · PASCAL-50S (5 references), MM pairsaccuracy (%): 66.8 -> 68.2Improved
use SPICE instead of METEOR on all PASCAL-50S pairsSPICE · PASCAL-50S (5 references), all pairsaccuracy (%): 80.8 -> 78.8Hurt
use SPICE instead of METEOR on human-vs-model pairsSPICE · PASCAL-50S (5 references), HM pairsaccuracy (%): 94.2 -> 87.5Hurt

Later papers that compared with this method

7 later accepted papers report SPICE as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy