1. Introduction

Portrait segmentation is to separate the portrait in a picture from the background to form different areas, which are distinguished using different tags.

This interface can be used to perform portrait segmentation on the part that includes a portrait on an input image. The expected segmentation result is that the portrait and the background are displayed in a separated manner.

2. Application Scenarios

Application scenarios: This algorithm is used to separate the portrait and the background in a picture and can be applied to scenarios such as background replacement and background blurring functions of APPs.

3. Capability Characteristics

  1. Rapidness: This algorithm is currently developed based on the deep neural network, to fully utilize the neural processing unit (NPU) of Huawei mobile phones to accelerate the neural network, achieving an acceleration rate of over 10 times.

  2. Compactness: By using this API, algorithm development time can be significantly reduced and the ROM space occupied by algorithm models is reduced, making your applications more compact.

4. Description

  1. Portrait segmentation applies to only pictures with portraits.

  2. Poor quality of a picture may affect the segmentation effect. User discretion is recommended.