Planning and Automatically Capturing the Desired Moment in Photography
Zhiyun Li, Ruoyun Li, Qingyun Li · 2024
The conventional practice of photography demands a considerable degree of patience and attentiveness, often requiring swift action to seize the perfect moment. In response to these challenges, our approach enables cameras to autonomously plan and capture the desired moments, significantly reducing the need for human intervention. The proposed technique utilizes voice or text descriptions to convey the user’s desired photographic outcome. These descriptions are transformed into encoded vectors, which serve as the foundation for the camera’s decision-making process. The camera continuously evaluates the similarity between the provided description and the ongoing preview image sequence, thereby facilitating the anticipation of the ideal photographic moment. Once the desired moment aligns with the description and occurs within the preview sequence, our system automatically captures the optimal shot with minimal human interaction.