Google Exploring "4D" Camera Tech & Stabilization Techniques

AH 2019 Google Pixel 3 AH NS camera 01

A newly-approved Google patent spotted by Android Headlines now seems to suggest that Google is looking to take its camera features to the next level with an invention it calls a 4D camera tracking system with optical image stabilization (OIS). The recently patented creation looks to adapt cameras away from the current methods used in image stabilization for both video and photographs while enabling more user-side features.

The company describes the invention as a camera system that utilizes light-field data to not only stabilize the image or video by generating predicted the camera’s ‘pathways’ using tracking across multiple selected pixels. That same feature can be used to both stabilize an image with a reduction to the amount of edge cropping required but can also better focus depth changes after a photo or image is captured. ‘Center of Perspective’ alterations are enabled post-capture as well.

What does this really mean?


Summarily, the technology found in Google’s latest patent may point to far more than better stabilization since it effectively can create ‘3D maps’ of each pixel. In that sense, it may be more closely described as a camera technology that generates 3D pixel maps from 2D snapshots accompanied by light-field data and motion path detection that add “extra dimensions” to the capture process itself.

That’s all assisted by a sensor that more accurately and comprehensively captures the light data, to begin with, making this as much a hardware solution as a software adjustment. All of that technology would be applied to generate a depth map for each of the keyframes and to offset initial camera motion while pixels are being captured.

That should mean anything from smoother single-frame or video captures for those who don’t necessarily have steady hands and the ability to change the perspective of the captured media on the fly based on the above-mentioned 3D maps. Conversely, it may also help to create even more accurate color and light balancing to offset uneven lighting.


Google undoubtedly has other uses planned for this technology well in terms of user-facing features.

A fitting terminology for the Pixel 4

The term “4D” in the context of the new invention may or may not be completely accurate but it would, in fact, make for a good marketing term if used in conjunction with one smartphone in particular. Namely, it wouldn’t be too surprising to see this feature find its way over to Google’s as-yet-unannounced Pixel 4 family of flagship smartphones.


Google Pixel handsets, with the exception of one or two flaws that have primarily been tied in with software, have consistently held out as among the best, if not the best, cameras available in a mobile device. So it would not only make sense from a marketing perspective to implement a “4D” camera on the Pixel 4 but as a way to possibly push smartphone camera technology forward too.

The feature would follow other innovative additions available only on the search giant’s devices, at least initially, such as its Night Sight feature that allows users to gain more detail even in extreme low-light settings. As with that technology, the introduction of Google’s latest patent into real-world smartphones would almost certainly make its way to other handsets through their respective OEMs’ own inventions and approximations or via ports.