Technical articles · 2025.04
Quality characteristics and verification of multimodal data for generative AI
생성형 AI 학습을 위한 멀티모달 데이터 품질특성 고찰 및 적용방안
Overview
The 2025 TTA Journal article explains representation, cross-modal correspondence and temporal/spatial alignment in multimodal training data. It applies these quality perspectives to tasks including audiovisual speech recognition, object/action classification, event/anomaly detection and image/video captioning.
Publication and version information
TTA Journal 218, March/April 2025, pp. 70–75. The month shown follows the publisher’s April 2025 index entry.
Version: TTA Journal 218, pp. 70–75