AI Paper Atlas

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

Measuring and Predicting Tag Importance for Image Retrieval

Shangwen Li, Sanjay Purushotham, Chen Chen, Yuzhuo Ren, C.‐C. Jay Kuo

TPAMI 2017 · Vision-language and multimodal › Human and embodied modeling · approach: Composed and text-based image retrieval · 34 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 structured model instead of Berg et al. for binary tag importancetag importance prediction · UIUCaccuracy: 82.0% -> 87.6%Improved
use continuous-valued instead of binary-valued relevance (subjective test)retrieval relevance function · COCO Scenehuman preference: 1176 out of 1500 or 78.4%Improved
add visual, semantic and context features to equal-importance baseline (ablation)structured prediction model · COCO, COCO Sceneprediction error: approximately 40% gainImproved
add visual features (ablation)tag importance model · COCO, COCO Sceneprediction error: approximately 28% reductionImproved
add semantic features (ablation)tag importance model · COCO, COCO Sceneprediction error: approximately 11% reductionImproved
add context features (ablation)tag importance model · COCO, COCO Sceneprediction error: approximately 11% reductionImproved
use ground truth bounding boxes instead of detected (ablation)tag importance model · COCOprediction error: approximately 3% performance loss from detection errorImproved
add visual features alone (ablation)tag importance model · UIUCprediction error: 1% performance loss vs baselineHurt
use continuous instead of binary predicted tag importancetag importance model · COCO and COCO Sceneprediction error: 48% and 45% loss for binaryImproved
use MIR/PCTI instead of visual baselineimage-to-image retrieval, K=50 · COCO and COCO SceneNDCG: approximately 14% and 10% gainImproved
use predicted continuous importance instead of binary-valuedMIR/PCTI vs MIR/PBTI, T2I K=50 · COCO and COCO SceneNDCG: 2% and 5%Improved
use MIR/PCTI instead of Traditional MIRT2I K=50 · COCO and COCO SceneNDCG: 4% and 11%Improved

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