PHELE: Completely Labelled Image Dataset for Physical Hazards of the Elderly Living Environment

The primary objective of the PHELE dataset is to facilitate the development of computer vision models and algorithms that can automatically detect potential hazards, which may pose a significant risk to the security and overall quality of life of the elderly. The dataset encompasses a diverse range of images, representing both indoor and outdoor settings commonly encountered in the living environments of elderly individuals. Each image in the dataset is accompanied by comprehensive labels and annotations. The labels denote different types of physical hazards that can exist in the living environment of the elderly, including but not limited to: Fire, Dog, Cat, Mouse, Chair,Hazardous electrical cords or wires, Lack of handrails or grab bars, Uneven or broken steps/staircases, Sharp edges or objects,Bed bugs, Mold/dampness, Peeling paint, Animal hair.

搜集汇总

背景与挑战

背景概述

PHELE数据集旨在帮助开发计算机视觉模型，自动检测老年人生活环境中的物理危险。它包含室内外场景的多样图像，每张图像都有完整的标签，标注了火灾、宠物、危险电线、缺少扶手、不平整台阶等多种危险类型。