BZ
Back to side projects

ML Engineer (10 Academy project)

Automated Storyboard Synthesis for Digital Advertising

ML solution that automated storyboard creation for digital ads by combining EDA on creative assets, YOLO object detection, and image generation.

Project gallery

Use the side arrows to browse; click an image for full size.

Cut storyboard production time from a week to a few hours, enabling faster delivery in a competitive market.

Feb 2024·1 min read
PythonYOLOPyTorchImage GenerationEDA

Automated Storyboard Synthesis for Digital Advertising

Storyboard production dropped from about a week to a few hours. A 10 Academy project that automates turning ad concepts and assets into storyboards for digital campaigns.

The problem

Agencies build storyboards by hand for every campaign concept: slow, repetitive, and a bottleneck in a market where turnaround speed wins pitches.

What I built

  • EDA on creative assets to understand the formats, objects, and layouts that make up real ad creatives.
  • YOLO object detection to identify and place creative elements from source assets.
  • An image-generation step that assembles concept frames into a coherent storyboard aligned to the brief.
  • An end-to-end pipeline that takes concept inputs and produces storyboard frames.

Result

What took roughly a week of manual work compressed into a few hours, letting the team iterate on more concepts per pitch. Full write-up on Medium.

Stack

Python, YOLO, PyTorch, image generation, exploratory data analysis.