Table 1 Targets of human data extraction in computer-vision papers and patents
From: Computer-vision research powers surveillance technology
Human body parts | Example: “The acquisition system may include a biometric sensor (for example, an electronic fingerprint sensor, or an optical eye scanner, or a camera arranged to acquire a portrait image of an authorized person’s face)” (Patent 71) |
These technologies most frequently target faces, including the detection of eyes, eye movement, faces, ‘suspicious’ facial expressions and facial recognition. Some targeted other body parts, typically to enable human activity recognition. | |
Human bodies | Example: “People monitoring in public areas, smart homes, urban traffic control, mobile application, and identity assessment for security and safety” (Paper 53) |
These technologies most frequently target humans in the midst of everyday activities (for example, walking, shopping and attending group events) for purposes including body detection, tracking and counting as well as security monitoring and human activity recognition. | |
Human spaces | Example: “A scene could be decomposed into a set of semantic objects” (accompanied by an example image taken inside an office) (Paper 40) |
These technologies most frequently target living spaces, personal and communal, such as people’s homes, offices, roads, town squares or borders. The purpose is often to identify unspecified objects in these spaces. Other specific purposes include modelling traffic and monitoring large border crossing areas. | |
Other socially salient human data | Example: “Free-hand human sketches [for example, of another person’s item of clothing] are used as queries to perform instance-level retrieval of images” (Paper 81) |
These technologies target data containing traces of the mental, economic, cultural, social status, identities, preferences or location details of humans, most frequently to narrow users’ search results. |
