Body Segmentation Technology

Body Segmentation Technology

Precise body segmentation SDK for mobile, desktop, and web applications.

Why Choose Banuba’s Body Segmentation Technology?

Why Choose Banuba’s Body Segmentation Technology?

  • High accuracy on bodies of different sizes and shapes
  • Works on multiple bodies in one frame/image
  • Reliable even in low-light conditions
  • Supports iOS, Android, Web, Desktop, and Unity
  • Easy integration thanks to npm/Maven/CocoaPods packages
  • Proven track record in photo and video apps

Use Cases for Human Body Segmentation

  • Video Communication

  • Photo Booths

  • Virtual Mirrors

  • Photo and Video Apps

Banuba’s Body Segmentation

  • How It Works
    1. The system takes the video frame from the feed/recording.
    2. Neural networks detect human body.
    3. The body is separated from the background and made ready for effects.

     
  • Real-Time

    Our Body Segmentation is based on the Fully Convolutional Neural Network. It classifies the input image pixels into a human silhouette or the background and builds a dynamic body mask. The technology lets you extract backgrounds or overlay an AR body effect in real-time and videos.

Banuba SDKs

Body segmentation is a part of Banuba’s SDKs (software development kits) – ready-made modules that can be quickly integrated into an app.

  • Face AR SDK for live video feed, Video Editor SDK to record, edit, and enhance TikTok-like content.
  • Augmented reality for Mobile, Web, Desktop, and Unity apps
  • Face detection & tracking, virtual makeup, face filters, beauty effects, and more
  • Live streaming, social media, photo/video editing apps, games, etc.
  • Unique video editing features like AI-generated subtitles, AI clipping, etc.
  • Runs on 90+% of smartphones
  • Low battery power consumption and heat

Banuba Face AR SDK

Powerful suite of augmented reality features that can be integrated in minutes. See the documentation for yourself.

Technical Features & System Requirements

logo-apple-3

iOS

From iOS 13.0
logo-android-4

Android

From Android 8.0 (API level 26)
img_TF_web-1

Web

All browsers
FAQ
  • Body segmentation means separating the human body from the rest of the image or video. It is a foundational technology for many other applications, including virtual try-on, virtual backgrounds, avatars, etc. 

  • Body segmentation technology uses trained neural networks to detect the presence of people in the image or video. These networks then “color” the necessary area, separating it from the rest of the image.

  • Banuba’s body segmentation technology can run on Web, Desktop, and Mobile platforms. This includes cross-platform mobile frameworks (e.g. React Native and Flutter) and Unity game engine. The large number of compatible technologies makes Banuba’s product convenient for almost any business interested in body segmentation.

  • A body segmentation model could be applied to one image several times to detect several people in one frame. This is a scalable approach, but it’s limited by the hardware. For example, on a desktop computer the performance would be better than on a smartphone.

  • The processing time for real-time body segmentation varies greatly depending on the specific neural network used, the hardware it runs on, the number of people in the frame, and other factors. Specific tests are needed to answer this question faithfully.

  • As body segmentation is foundational for many filters and effects, it is most commonly used in social media and video communication, as well as video editing. Other verticals, e.g. marketing, eLearning, and eCommerce, can also use this technology to great effect.

Get Free Trial