Photography/ Video Bokeh

TetrasMobile Video
Image SDK Solution

Tetras.AI's intelligent imaging software solution delves deep into the entire process of phone photography, from sensors to ISP, to the entire application. It is an end-to-end computational photography solution with a product line that covers multiple key scenarios such as night scenes, super resolution, portraits, and videos, assisting phone manufacturers to comprehensively improve photo/video imaging quality and enhance users' photographing/video shooting experience.

Product Introduction

Our bokeh product supports both single and dual camera bokeh in photography. The depth of photos taken by single and dual cameras is estimated through high-quality and high-performance semantic segmentation and depth estimation engines. The subject is highlighted through image/video segmentation, and the stereoscopic impression of the photo is enhanced through progressive bokeh processing. This allows phones to achieve a large aperture effect similar to that of a professional camera. Our product enhances the professional portrait photography experience and is widely used in portrait mode and large aperture mode. More than 200 models incorporate this product, benefiting hundreds of millions of consumers.

We also support portrait bokeh, fine edge segmentation, and real-time running speed of the video in movie mode. This improves the overall video experience as if watching a major motion picture.

Supported Platform

At present, Tetras.AI's bokeh product has been adapted for use on multiple platforms such as Qualcomm and MTK. Combined with the multiple computing units of these platforms, our product is able to run smoothly, integrate into, and be optimally tuned on low- to high-end platforms.

Product Advantages

We use a precision depth-estimation engine to render a progressive bokeh effect. A precision AI segmentation model is also used to ensure clearer and more accurate subject/object edge processing. In addition, our bokeh product adapts to multiple units and platforms, providing users with a smooth photography experience.

Key Technologies

Depth Estimation Engine

The depth of photos taken by single and dual cameras is estimated through high-quality and high-performance semantic segmentation and depth estimation engines.
AI Segmentation Model

The subject is highlighted through image/video segmentation, and the stereoscopic impression of the photo is enhanced through progressive bokeh processing.
The Dual-Camera Parallax Calculation

Dual-Camera parallax is calculated using the baseline size and imaging difference between the dual cameras and AI technology.

Application Scenarios

Portrait Mode Preview
Portrait Mode Photography
Video Portrait Bokeh