AI Paper Atlas

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

Interleaved Text/Image Deep Mining on a Large-Scale Radiology Database for Automated Image Interpretation

Hoo-Chang Shin, Le Lu, Lauren Kim, Ari Seff, Jianhua Yao, Ronald M. Summers

JMLR 2016 · Vision-language and multimodal › Human and embodied modeling · approach: Neural image captioning · 37 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 VGG-19 instead of AlexNet (ablation)CNN image-to-topic classifier · document-level topicstop-1 accuracy: 0.607 -> 0.658Improved
use VGG-16 instead of AlexNet (ablation)CNN image-to-topic classifier · document-level sub-topicstop-1 accuracy: 0.33 -> 0.54Improved
use VGG-16 instead of AlexNet (ablation)CNN image-to-topic classifier · sentence-level topicstop-1 accuracy: 0.48 -> 0.50Improved
add three conv layers (VGG-19 vs VGG-16) (ablation)VGG-16 · document-level topicstop-5 accuracy: 0.93 -> 0.95Improved
add three conv layers (VGG-19 vs VGG-16) (ablation)VGG-16 · document-level sub-topicstop-1 accuracy: 0.54 -> 0.54No clear effect
use deeper VGG-19 instead of AlexNet (training cost) (ablation)AlexNet · training time to 70,000 iterationstraining time: 4 hours 35 mins -> 4 days 40 minsHurt
initialize from ImageNet pre-trained weights instead of random (ablation)VGG CNN · image-to-topic classificationtest accuracyImproved
fine-tune from related higher-level topic CNN instead of ImageNet model (ablation)CNN for sub-topics · document-level sub-topictest accuracyImproved
deeper VGG instead of AlexNet only when pre-trained on ImageNet (ablation)VGG-Net · image-to-topic classificationtest accuracyMixed
add OpenI biomedical articles to word2vec training corpus (ablation)word2vec on radiology reports · word vector qualitycosine similarity of neighbors: ~1 billion -> ~1.2 billion wordsImproved
restrict labels to specific disease terms (UMLS T047 with negation) instead of loose topic labelsimage labeling · dataset coveragefraction of images with labels: about 90% of images lostHurt
restrict LDA to reports with associated images (ablation)document-level LDA · topic miningnumber of meaningful topics: 20 topicsNo clear effect

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