In Which The Author Explores Machine Learning for Depth Extraction

Tags

Machine Learning

Author

Ganz

Date

September 28, 2024

Post No.

Chroma 0301

URL

https://www.linkedin.com/posts/heyganz_alphachromatica-chroma-heyganz-activity-7242608394241499136-V9zG?utm_source=share&utm_medium=member_desktop

Machine Learning is no longer just a concept for the future - it’s already changing how we work today. Recently, I ran a quick test to explore how machine learning can be applied to produce physically plausible depth maps.

Take a look at these images :

1️⃣ Top: The Last Scene from the Oscar winning movie from 2015, Birdman

2️⃣ Middle: Using a machine learning, I extracted depth data from the footage, creating a depth map that reveals the spatial layers of the scene.

3️⃣ Bottom: With the depth map, I isolated the foreground from the midground and background

This is just one small example of how AI can simplify tasks that were previously labor-intensive. The potential is huge - not just for automating repetitive work, but for giving us more room to experiment creatively.

I’ve been testing tools like this to find where they fit into our everyday workflows at Alpha Chromatica. It’s not about replacing manual techniques, but about seeing where technology can speed things up or offer new creative and technical possibilities.