AI Paper Atlas

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

BreakingNews: Article Annotation by Image and Text Processing

Arnau Ramisa A, Fei Yan, Francesc Moreno-Noguer, Krystian Mikolajczyk

TPAMI 2017 · Vision-language and multimodal › Human and embodied modeling · approach: Neural image captioning · 116 citations (OpenAlex)

Read the paper

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 CNN on full Word2Vec matrix instead of SVM on best shallow featureSVM BoW TF-IDF · Source detectionbalanced accuracy (%): 72.6 -> 80.7Improved
use full BoW instead of truncated 5000-d BoW (ablation)SVM · Source detectionaccuracy (%): 70.8 -> 72.9Improved
use mean of Word2Vec instead of max (ablation)SVM · Source detectionaccuracy (%): 57.3 -> 68.6Improved
combine VGG19 and Places features instead of VGG19 alone (ablation)CCA W2V mean · Article illustrationmedian rank: 499 -> 415Improved
use CNN with full Word2Vec matrix instead of CCA with W2V meanCCA W2V mean + VGG19+Places · Article illustrationmedian rank: 415 -> 405No clear effect
replace article image with related Internet images (ablation)CCA W2V mean + VGG19 · Article illustrationmedian rank: 499 -> 161Improved
add original image caption to visual representation (ablation)CCA W2V mean + VGG19 · Article illustrationmedian rank: 499 -> 7Improved
anonymize proper nouns in the added caption (ablation)CCA with captions · Article illustrationmedian rank: 7 -> 42Hurt
use CNN instead of SVR with W2V meanSVR W2V mean · Popularity predictionmedian l1 error: 3.63 -> 1.09Improved
use CNN instead of SVR on mean l1 errorSVR W2V mean · Popularity predictionmean l1 error: 69.75 -> 70.72No clear effect
use GCD loss instead of Euclidean loss (ablation)CNN · Geolocation predictionmedian GCD (thousand km): 1.42 -> 0.90Improved
train geolocation and popularity jointly (multitask) (ablation)CNN single-task · Geolocation + popularitymedian GCD / median l1: 0.90 -> 1.17; 1.09 -> 1.68Hurt

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