Google now open sources its Pixel 2’s portrait mode AI

Google’s Pixel 2 and Pixel 2 XL take exemplary portrait mode photos.

Many believe that the phone’s photos are even better than the iPhone X and Samsung’s line of flagship handsets. Now Google has openly sourced its portrait mode artificial intelligence (AI).

Portrait mode uses Google’s DeepLab-v3+ system. This gives each pixel a label to identify objects in the background and foreground of an image.

“Semantic image segmentation, the task of assigning a semantic label, such as ‘road,’ ‘sky,’ ‘person,’ and ‘dog,’ to every pixel in an image enables numerous new applications, such as the synthetic shallow depth-of-field effect shipped in the portrait mode of the Pixel 2 and Pixel 2 XL smartphones and mobile real-time video segmentation,” writes Google software engineers Liang-Chieh Chen and Yukun Zhu, in a recent research blog.

Google built DeepLab-v3+ using convolutional neural networks (CNNs), a machine learning method that analyzes visual data.

DeepLab-v3+ will be available to developers to integrate into their own apps.

“Since the first incarnation of our DeepLab model three years ago, improved CNN feature extractors, better object scale modeling, careful assimilation of contextual information, improved training procedures, and increasingly powerful hardware and software have led to improvements with DeepLab-v2 and DeepLab-v3,” according to the blog post.

It’s unlikely Android manufacturers like Samsung or Huawei will copy the Pixel 2’s portrait mode. Most manufacturers already have their own portrait mode and even though some are better than others, it’s unlikely the big players in the Android space will take advantage of Google’s portrait technology.

That said, it will be interesting to see what apps adopt Google’s portrait mode artificial intelligence.

You can find the full research blog post here.

Source: Google