SizeUp AI

Inspiration

I built SizeUp AI after being inspired by the time-consuming manual resizing for various social media platforms.

What it does

It's an Adobe Express add-on that uses AI to save hours by effectively re-cropping and adjusting text.

How we built it

In order to facilitate user interaction, I developed SizeUp AI as a two-part system, first is an Adobe Express Add-on frontend (HTML/CSS/JS with Express SDK). Secondly we have Python backend (Flask, Pillow, and computer vision libraries) which is connected to this frontend through API calls. To maintain content focus, the backend uses AI to intelligently re-crop images and recommend text changes. It is a multi-platform content workflow solution that is automated.

Challenges we faced

it was a challenge connecting JavaScript add-on frontend with the Python AI backend, particularly with secure data transfer. Improving the AI for smart re-cropping and text adjustments within the hackathon's scope was also a major obstacle.

Accomplishments that we're proud of

By automating content adaptation, SizeUp AI guarantees flawless, consistent posts on all platforms.

What we learned

I gained valuable insights into image processing, computer vision, and leveraging AI to enhance creative workflows.

What's next for SizeUp AI

For SizeUp AI, I plan to improve its core AI. This will allow for more advanced layout changes and smart content adjustments, going beyond simple re-cropping. I also want to expand optimization to video clips and give users more control with custom presets.

Built With

• os
• python
• subprocess
• sys
• vscode
• webbrowser