Summary

This site provides access to ImageSense, an integrated intelligence engine for more than hundred million open-source ground-level images drawn across the web, accompanied by extracted knowledge tags, geotags and range of other valuable metadata. Explore a detailed index of tags, keywords, and other features for up-to-the-minute information on places of interest around the world. Use this data to draw digital footprints for location, build functional maps, improve population estimates, identify geographic points of interest, validate intelligence, and track events as they happen.

Objectives

  • Spatially explicit damage assessment of critical infrastructure
  • Situational awareness to address complex national security challenges
  • Place and scene understanding of complex points of interest
  • Functional mapping of the planet using crowdsourced ground photos
  • Data gathering and conflation at scale for landmark classification
  • Geographical, cultural, and political attribution
  • Coordination and planning of post-disaster recovery efforts
  • Extract multi-level features from images and annotate them in real-time
  • Improve object and scene detection and downstream analysis

Features

eye image

Object Identification

Recognize and classify objects, features, scenes, and points of interest via various deep learning solutions
link image

Comprehensive Place Analytics

Automatically extract multi-level attributes and features from images. Perform range-of-place analytics for knowledge discovery, including functional categorization, recognition of geographical attributes, higher-order information retrieval.
text image

Joint Image and Text Representation Learning

Perform joint visual and textual understanding and build downstream analytics solutions, exploiting multimodal information. Generate textual interpretations and actionable intelligence from different types of scene images
lock image

Higher Order Intelligence

Multi-level deep knowledge extraction driven complex intelligence engine, moving beyond only observable lower-order intelligence
globe image

Global Reach

Explore more than 100 million geospatially coded images from multiple sources
aperture image

Image to Text

Uncover geographic tags and contextual information hidden within images
magnifier image

Searchable Data Cloud

Discover relevant images based on tags, points of interest, features, and encoded scene information. Browse across more than 350 indexed tags and keywords
map image

Geolocated

Automatically detect, link, and catalog images by locations and geographic coordinates
download image

Software as a Service

Integrate with third-party platforms and APIs for added value on data and analytical insights
brain image

Explainable AI

Get an edge by employing explainable artificial intelligence solutions for place and scene analytics, stepping away from uninterpretable black box machine-learning models

Scientific Artifacts

YearCitation
2021

Gurav, R., De, D., Thakur, G., & Fan, J. (2021). Conflation of Geospatial POI Data and Ground-level Imagery via Link Prediction on Joint Semantic Graph. In Proceedings of the 4th ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery (GeoAI'21), pp. 5-8. https://doi.org/10.1145/3486635.3491068 
2022[Under review] De, D., & Thakur, G. (2022). ImageSense: Extraction of Higher-order Attributes from Few-shot Ground-level Imagery of Places, Scenes. Under review at IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 
2021Thakur, G., Sims, K., Rittmaier, C., Bentley, J., De, D., Fan, J., Liu, T., Palumbo, R., McGaha, J., Nugent, P., & Eaton, B. (2021). Accelerated Assessment of Critical Infrastructure in Aiding Recovery Efforts During Natural and Human-made Disaster. In Proceedings of the 29th International Conference on Advances in Geographic Information Systems (SIGSPATIAL'21). pp. 195-206. https://doi.org/10.1145/3474717.3483947
2019 Chouhan, S., & Thakur, G. (2019). Contextual Deep Learning Framework for Joint Object Detection and Scene Classification of Ground Level Photos. In: Northern Lights Deep Learning Workshop. 2019. 

Contact Us

Debraj De

Location Intelligence Group
Human Dynamics Section,
Geospatial Science and Human Security Division
National Security and Sciences Directorate
Oak Ridge National Laboratory

[email protected]

Gautam Thakur

Location Intelligence Group
Human Dynamics Section,
Geospatial Science and Human Security Division
National Security and Sciences Directorate
Oak Ridge National Laboratory

[email protected]