ForestPersonsIR: Infrared Images for Missing Person Detection in Forests

ForestPersonsIR is a large-scale dataset designed for missing person detection in forested environments under search-and-rescue scenarios. It contains images captured using IR sensors at ground and low-altitude perspectives, simulating realistic field conditions with varied poses and visibility levels. The dataset was created by ETRI and is associated with an ICLR 2026 paper.

Use Cases

• Training object detection models for missing persons based on infrared imagery.
• Evaluating model robustness under varied visibility levels (20, 40, 70, 100) mentioned in the description.
• Developing algorithms for person detection in forested environments from ground and low-altitude perspectives.
• Benchmarking performance on varied human poses (standing, sitting, lying) in natural settings.

Strengths

• Simulates realistic search-and-rescue field conditions with varied poses and visibility levels.
• Images captured from multiple perspectives (ground and low-altitude).
• Associated with a peer-reviewed ICLR 2026 paper, suggesting academic validation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license are unknown, which may limit suitability assessment.

Provenance

Source

ETRI (Electronics and Telecommunications Research Institute).

Collection Method

Images captured using IR sensors in simulated forest environments.

Freshness

Last updated 2026-02-09 04:41:52; freshness should be verified.

Geography

Forest environments; specific location is unknown.