AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Incidents1M: A Large-Scale Dataset of Images With Natural Disasters, Damage, and Incidents

Ethan Weber, Dim P. Papadopoulos, Àgata Lapedriza, Ferda Ofli, Muhammad Imran, Antonio B. Torralba

TPAMI 2022 · Detection and segmentation › Open-set and few-shot detection · approach: Remote sensing segmentation and change · 66 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
add class-negative labels to the training loss (ablation)Incidents1M Model (ResNet-50) · Incidents1M test setincident mAP: 62.83 -> 67.19Improved
add class-negative labels to the training loss (ablation)Incidents1M Model (ResNet-50) · Incidents1M test setplace mAP: 63.15 -> 62.94No clear effect
train on 100% instead of 25% of Incidents1M (ablation)Incidents1M Model · Incidents1M test setincident mAP: 64.01 -> 67.19Improved
train on 100% instead of 25% of Incidents1M (ablation)Incidents1M Model · Incidents1M test setplace mAP: 59.68 -> 62.94Improved
train on 50% instead of 25% of the data (ablation)Incidents1M Model · Incidents1M test setincident mAP: 64.01 -> 65.59Improved
use multi-label Incidents1M data (about 2x images) instead of the original Incidents datasetResNet-50 same architecture and loss · Incidents1M test setincident mAP: 63.89 -> 67.19Improved
use multi-label Incidents1M data (about 2x images) instead of the original Incidents datasetResNet-50 same architecture and loss · Incidents1M test setplace mAP: 60.33 -> 62.94Improved
use only 25% of Incidents1M (446,788 images) versus original Incidents trainingIncidents1M Model · Incidents1M test setincident mAP: 63.89 vs 64.01No clear effect
filter Twitter images with the Incidents1M Model before retrievalTwitter earthquake events · Twitter earthquake eventsaverage precision (%): 15.87 -> 77.71Improved
filter Twitter images with the Incidents1M Model before retrievalTwitter flood events · Twitter flood eventsaverage precision (%): 35.68 -> 92.96Improved
evaluate on geotagged Twitter images instead of the Incidents1M test setIncidents1M Model · Geographical bias testmAP: 69.94 -> 53.46 (best continent EU)Hurt
filter tweets by incident images for wildfire correlation with eventsIncidents1M Model temporal monitoring · Twitter vs NOAA eventsIoUMixed

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