Volume is Using Machine Learning to Transform 2D Images to 3D
Overview
- Volume is a tool for transforming 2D images to 3D
- The model requires tons and tons of data to learn and improve
- The tool is open to the public as well
- Check out the video below to get an idea of how the technology works
Introduction
Remember those 3D masks in the Mission Impossible movies? Using just a 2D image, they were able to effortlessly create them from scratch. Well, thanks to deep learning, we are not too far away from fiction becoming reality.
Source: dornob
Volume is a tool for reconstructing any 2D image (or video) in a 3 dimensional space. To build a model that understands human movements requires a gigantic amount of data. According to one of the co-developers, Or Fleisher, they are building a neural network classification model to identify the elements in an image that are the most dominant.
Volume has opened it’s tool to the general public – so anyone can go and upload their image(s) to the site. You can try to do it yourself here.
The initial resulting images range from impressive to downright funny. As you zoom into the image, you can see how the pixels are broken down into different parts. It makes for fascinating viewing. But as users keep uploading images, Volume’s model keeps learning with the increase in the training data.
Since this concept is still in it’s teething stages, the developers are working to continuously improve the understanding and accuracy of the model.
Check out the below incredible video to see the tool in action:
Our take on this
While the model and the released tool are still in early development, one can imagine this technology being applied in the gaming industry. EVen video editors, currently using an amalgamation of tools, could find this to be an immensely helpful addition.
A similar AI was revealed a few months back by researchers in the UK which converted 2D images of faces to 3D. But Volume does a lot more than just facial recognition. It looks at the entire image and attempts to convert it to 3D. That, as we mentioned, takes a massive amount of data to train and improve. It’s an audacious project being attempted but it opens the door for others to try it out as well.
Subscribe to AVBytes here to get regular data science, machine learning and AI updates in your inbox!
Senior Editor at Analytics Vidhya. Data visualization practitioner who loves reading and delving deeper into the data science and machine learning arts. Always looking for new ways to improve processes using ML and AI.
