Quick Tip: Using Machine Learning on portrait-oriented still images to fill a 16:9 frame
Machine Learning tools can do some amazing things. They can even fit a vertical image into a wider 16:9 hole. I’m not talking about increasing the image’s resolution so the image tolerates a 200% or 300% scaling operation.
I’m talking about extending the rest of the vertical image to fill in the horizontal empty space with new unseen image detail!
Think: Hand-painted set extensions (like a matte painting artist) but with a computer creating the set extensions with a few clicks of a button.
In this Mixing Light Short, we explore using Machine Learning (ML) to solve a common editorial aesthetic problem, fitting vertical 9:16 images in a 16:9 horizontal frame. We’ll do this in just a few clicks using Runway’s Infinite Image machine learning tool – a tool that extends your image based on its visual content.
I want you to know where this technology is taking us and provide an example you can use with your clients today. Of course, when uploading client images to these services, be sure you have your client’s approval to avoid copyright issues.
Mixing Light members, log in to view the discount code for RunwayML – it’s above the video.
- Visit RunwayML.com – Runway currenly has over 30 ML-based tools that are interesting to experiment with. I encourage you to check them out. You’ll notice that some of these tools are also showing up in plugins for DaVinci Resolve and other post-production tools (such as depth matte extraction, removing/healing parts of an image, and super slow motion).
Related Mixing Light Insights
- A Filmmakers Introduction to Artificial Intelligence and Machine Learning – This Insight is an introduction to those terms (for filmmakers). Learn how to put those buzzwords in context; and how they inform you about the technology being discussed.
- The Future of Work: Can computers be creative (and take your career)? – Creativity makes us human. What is the place of AI in this area? If computers become too human, will we become just machines?