Getting Crevices, Cracks, and Grooves in Line: Anomaly Categorization for AQC Judgment Models

Abstract

The production of high-end manufactured products requires expensive human aesthetic quality control (AQC) in the form of e.g. visual inspection at multiple stages during production. Current standards for aesthetic quality control focus on the process, i.e. identifying the source of anomalies and lack an observer-oriented classification (describing the perceived appearance). We found a need for a perceptual categorization of anomalies that can help assessors making quality judgments and we present a judgment model for AQC. To this end, we studied visual inspection task flows, and processes especially around limit samples at Bang & Olufsen. Based on perceptual responses in the human visual system, we propose a categorization consisting of open and closed shapes. The latter contains areas, lines, alignment, and constellations. The categorization can help in automating AQC using computer vision.