2020/06/01 by José David Fernández-Rodríguez, Rodríguez, Jose David Fernández, David Daniel Albarracín-Molina +4
Computer Science · Engineering · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Human Motion and Animation #Human Pose and Action Recognition #Video Analysis and Summarization
paper · pdf · doi:10.48550/arxiv.2006.00904
openalex publication_date 2020/06/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Oftentimes TV producers of motor-racing programs overlay visual and textual\nmedia to provide on-screen context about drivers, such as a driver's name,\nposition or photo. Typically this is accomplished by a human producer who\nvisually identifies the drivers on screen, manually toggling the contextual\nmedia associated to each one and coordinating with cameramen and other TV\nproducers to keep the racer in the shot while the contextual media is on\nscreen. This labor-intensive and highly dedicated process is mostly suited to\nstatic overlays and makes it difficult to overlay contextual information about\nmany drivers at the same time in short shots. This paper presents a system that\nlargely automates these tasks and enables dynamic overlays using deep learning\nto track the drivers as they move on screen, without human intervention. This\nsystem is not merely theoretical, but an implementation has already been\ndeployed during live races by a TV production company at Formula E races. We\npresent the challenges faced during the implementation and discuss the\nimplications. Additionally, we cover future applications and roadmap of this\nnew technological development.\n