Photosynth Returns as a Feature in Microsoft’s Pix App

Microsoft Photosynth was first introduced in 2008, allowing users to stitch together 3D panoramas called “synths” from their still photos. The service was shut down in February, but now the software maker has brought back some of the underlying technology as a new feature in the Microsoft Pix app for iOS.

Pix is a free photo-taking and video-capture app for Apple iPhones and iPads that employs artificial intelligence techniques to automatically improve image quality, helping users snap Instagram-worthy photos. The latest update revives Photosynth as a new mode that allows users to take larger, more immersive panoramas.

Shooting a panorama in most standard photo apps generally requires using a smartphone camera to cut a narrow horizontal swath on one direction across the user’s field of view. It’s a good app for capturing distant skylines and other scenes where height isn’t much of a factor, but it falls short for users who may want to capture the depths of a canyon and the dramatic skies overhead.

The new feature in Pix borrows some of the image-composition functionality from its forbear and the app’s own auto-enhancement capabilities to capture an entire scene. Users can now point their cameras in practically any direction (but not too fast), sweeping back and forth, skyward and toward the ground, for fuller view of one’s surroundings.

Another feature, Pix Comix, uses AI to create comics from video footage.

“Pix uses its deep learning model to score and select three high-quality frames, searching specifically for non-blurry photos, faces with eyes open and interesting scenes,” wrote Microsoft marketing communications manager Nicky Budd-Thanos in a Dec. 20 blog post.

“Pix Comix performs this AI processing on device, selecting the best frames and formatting them into a comic strip. From there, you can add and edit speech bubbles that can be moved, rotated and resized to tell your own story.”

Google, meanwhile, is taking AI-enabled photography into a more subjective realm.

On Dec. 18, Google introduced an AI model called Neural Image Assessment, or NIMA for short, that can pick out pictures that attract attention. NIMA employs a machine learning technology called convolutional neural networks to predict which images a typical user would consider attractive and nice to look at.

In future, the technology could be used to automatically enhance photos in ways that increase the likelihood that they catch a person’s eye. “Our proposed network can be used to not only score images reliably and with high correlation to human perception, but also it is useful for a variety of labor intensive and subjective tasks such as intelligent photo editing, optimizing visual quality for increased user engagement, or minimizing perceived visual errors in an imaging pipeline,” explained Google software engineer Hossein Talebi and Google Research scientist Peyman Milanfar, in a Dec. 18 announcement.