소비자 참여행동에 영향을 미치는 Short-form 영상의 요소 : 컴퓨터 비전 알고리즘을 활용한 이미지 분석

초록

Recent growth in social networking services (SNS) and mobile-centric media consumption has ledto an explosive increase in user-generated content data. Short-form video, in particular, delivers vivid visualexperiences within a brief time span and has emerged as a media format well aligned with contemporarypatterns of information consumption. As short-form content proliferates across platforms, its role hasexpanded beyond individual expression to encompass firm value creation and marketing strategy. Becauseshort-form videos are typically distributed via recommendation algorithms, viewers must decide whether toengage with content under severe time constraints, making visual cues more consequential than in manyother media. Accordingly, this study decomposes short-form videos into frame-level images and examineshow visual attributes shape consumer engagement behaviors. Focusing on YouTube Shorts, we applycomputer-vision–based image analytics to convert videos into frame sequences and extract visual features—including aesthetics, quality, color characteristics, and detected objects—to test their effects on engagementoutcomes and to identify when these features exert the strongest influence over the course of a video. Thisresearch is among the first to analyze short-form social media content at the frame level, demonstratingboth the importance of frame-specific visual elements and the temporal loci at which their effects emerge. The findings provide a theoretical foundation and future directions for content automation grounded incomputational aesthetics, as well as optimization strategies to predict and enhance consumer responses inalgorithmically curated media environments.

키워드