https://umd.zoom.us/my/
In recent years, advancements in camera technology have democratized visual creativity, allowing individuals to document experiences and express themselves through video. However, video editing remains a complex, time-consuming task requiring both technical skill and creative intuition, limiting accessibility for non-professionals. While AI-powered tools have automated many aspects of editing, such as scene detection and cut suggestions, they still struggle with a critical element: understanding context. Context, comprising the broader environment, circumstances, and intent behind the video, is key to making informed editing decisions that enhance viewer experience.
Our work seeks to bridge the gap between AI automation and creative video editing by developing algorithms that incorporate contextual awareness into the editing process. Our research contributes to several key stages of post-production and focuses on three key areas including (1) context-based image editing, which aims to refine visual elements at the frame level; (2) context-aware visual effects, where transitions, camera motions and other effects are adapted to the narrative and creator’s intent for the content; and (3) viewer reception analysis, where audience engagement data can potentially be leveraged for iterative editing decisions. By embedding contextual understanding into AI-driven editing tools, we aim to streamline and enrich various stages of post-production, enabling the creation of personalized, high-quality video content that resonates with diverse audiences.
Pooja Guhan is a PhD student in Computer Science at the University of Maryland (UMD), College Park, where she is advised by Prof. Dinesh Manocha. She completed her Master’s in Computer Science from UMD in 2021 and her Bachelor’s in Electronics and Communication Engineering from IIIT Hyderabad in 2019. Her research interests include computer vision, deep learning, and affective computing. Currently, her work focuses on multimedia technologies, with a particular emphasis on computational video editing.