Enhancing Instream Shoppable Brand and Product Detection in Broadcast, OTT, and VOD Content through Multi-Model Object Detection and Real-Time SCTE-35/SEI/VMAP Integration 

The increasing adoption of shoppable video content in Broadcast, OTT, and VOD has transformed how consumers engage with digital media. Traditional advertising models struggle to create seamless, interactive, and real-time brand engagement, leading to a need for AI-driven solutions. Emerging technologies now enable real-time product detection, metadata embedding, and interactive instream commerce integration, making video content viewable and instantly shoppable. The paper introduces an AI-based platform that instantly detects brands and products in video streams before implementing metadata structures for interactive shopping platforms. By leveraging deep learning-based object detection models and metadata signaling protocols, the framework ensures synchronized, real-time engagement between consumers and brands, unlocking new monetization opportunities for advertisers.

Multiple AI detection engines under this framework based on YOLO, Mask R-CNN, ResNet, and MobileNet SSD precisely identify products and brand placements inside real-time and pre-recorded content. The detections from the system link up with SCTE-35 (live streams), SEI (compressed video streams), and VMAP (VOD-based ad scheduling), which enable precise frame-level interactivity for product engagement. Based on AI technology, the system allows moment-by-moment brand interactions and stream-based e-commerce capabilities, and advertisers can use it for content monetization. The framework enhances content creator and advertiser success while delivering improved consumer engagement since it automates product detection combined with advertisement synchronization features. This framework redefines media commerce by combining AI-based detection, metadata injection, and HTML-based interactivity, allowing for scalable, real-time, and seamless brand engagement in video content.

Chaitanya Mahanthi | Google, YouTube | New York, N.Y., United States