Robot Head Face Tracking Using Depth Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional human face detecting and tracking methods for robots suffer from high detection errors, leading to inaccurate rotation control of robot heads, as they often misidentify non-human faces as human faces.

Innovation Solution

A face detecting and tracking method that involves depth detection, determining face candidate areas, performing face detection using the Viola-Jones algorithm, and tracking faces using a compressive tracking algorithm, ensuring accurate face detection and smooth robot head rotation by centering the tracking box within the image frame.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional human face detecting and tracking methods are used, then the robot can detect and track faces, but the detection accuracy is low and non-human faces are misidentified as human faces

Engineering Contradiction:
Improveface detection accuracyVSAvoiddetection error rate
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent introduces depth information as an additional dimension to the traditional 2D image data. By obtaining depth values for each pixel and determining face candidate areas based on depth information, the system creates a 3D-aware detection framework that effectively filters out false positives from non-human faces while maintaining accurate human face detection.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent changes the detection parameters by incorporating depth values alongside color and texture information. The face detection process uses multiple parameters including depth, color, and texture to comprehensively identify human faces, thereby improving detection accuracy and reducing misidentification of non-human faces.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If face detection accuracy is improved using multiple parameters, then detection reliability increases, but the computational complexity and processing time increase

Engineering Contradiction:
Improvedetection accuracyVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by first obtaining depth information and determining face candidate areas before conducting full face detection. This preliminary filtering step reduces the search space and allows the system to focus computational resources on potential face regions, thereby managing complexity while maintaining high detection accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the detection process into distinct stages: depth-based candidate area determination, face detection within candidate areas, and tracking. This segmentation allows each stage to be optimized independently, managing overall system complexity while achieving high reliability through cumulative processing.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS9892312B1Face detecting and tracking method and device and method and system for controlling rotation of robot head
Publication Date: 2018.02.13 UBTECH ROBOTICS CORP LTD
  • US9892312B1 patent drawing
  • US9892312B1 patent drawing
  • US9892312B1 patent drawing

AI summary

A face detecting and tracking method includes: acquiring an image and performing a depth detection for the image to obtain a depth value of each pixel of the image; determining one or more face candidate areas based on depth value of each pixel of the image of current frame; performing a face detection to the one or more face candidate areas to determine one or more face boxes of the image of current frame; and determining a tracking box of the image of current frame based on the one or more face boxes and a tracked face box, and tracking the face in the tracking box of the image of current frame.