Last year, Google managed to surprise everyone by introducing a very impressive Portrait mode on the Pixel 2 despite the handset coming with only a single 12MP shooter at its rear. Given that the likes of Apple, Samsung, and Huawei had to use a dual-camera setup for this, Google’s approach was not only unique, the results were surprisingly even better than what some of the phones with a dual-camera setup could capture.

Now, in a move that will surprise a lot of folks, Google has gone ahead and made the tech behind the Portrait mode on the Pixel 2 open source.

The Portrait mode on the Pixel 2 is powered by Google’s DeepLab model, the first incarnation of which was released over three years ago. Since then, Google has improved DeepLab by taking better advantage of the powerful hardware and software integration.

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. With DeepLab-v3+, we extend DeepLab-v3 by adding a simple yet effective decoder module to refine the segmentation results especially along object boundaries.

The above details might seem too complicated but from a layman’s viewpoint, it means that other OEMs can take Google’s work and implement it in their future smartphones. This move from Google also shows just how confident it is about its technology. I cannot think of any other OEM but Google who would have gone ahead and open sourced such a key technology that is a part of its phone.