Mobile Device 3D Face Model Generation Using Inertial Sensor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating 3D face models using mobile devices face challenges in capturing clear multi-view images due to camera motion causing blurry images, lower image quality in video mode, and the need for manual adjustment and instruction, limiting their effectiveness for general users.
Innovation Solution
An apparatus and method that utilizes an inertial sensor and front-facing camera in a mobile device to automatically capture still images from multiple directions, adjusting camera properties for optimal feature point detection and allowing users to adjust the device's position and orientation without additional instruction, incorporating a multi-view image capturer and 3D model generator to create a precise 3D face model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the user moves the camera to capture images from different angles, then multi-view images can be obtained, but the camera motion causes the images to be blurry
Solution Approach 1:
The system performs preliminary actions by detecting face regions in advance before capturing images. The mobile device captures a sequence of images, detects face regions in each image, and uses these detections to guide subsequent capture decisions, ensuring clear images are obtained before motion occurs
Solution Approach 2:
The system implements feedback by continuously detecting face regions in captured images and using this information to control further capture. The face region detection results feed back to the capture control mechanism, allowing the system to adjust capture timing and positioning based on actual face detection outcomes
2Ease of operation
If the user captures images in video mode to allow continuous capture from different angles, then capture convenience is improved, but image quality is lower than still-image mode
Solution Approach 1:
The system uses partial action by capturing a limited sequence of images rather than continuous video. It captures N images in sequence, processes them to identify face regions, and selects only the necessary images for 3D model generation, avoiding the quality loss of full video mode while maintaining operational simplicity
Solution Approach 2:
The system changes capture parameters dynamically based on face detection results. It adjusts capture timing, positioning, and selection criteria according to detected face regions, transforming the capture process from a fixed video mode to an adaptive still-image sequence that maintains high quality
3Measurement precision
If the user presses a button to capture images manually, then capture control is precise, but the mobile device shakes and image quality is degraded
Solution Approach 1:
The system implements self-service by automatically controlling image capture based on face region detection. The mobile device itself detects face regions in captured images and autonomously determines when and where to capture next, eliminating the need for manual button pressing and the associated device shaking
Solution Approach 2:
The system replaces the mechanical action of manual button pressing with an automated detection-and-capture mechanism. Face region detection algorithms substitute for manual control, using computational methods to determine capture timing and positioning without physical user interaction that causes device movement
4Loss of time
If the user adjusts position and orientation at each time point using timer capture, then capture timing is controlled, but the user cannot have sufficient time to adjust the mobile device
Solution Approach 1:
The system implements dynamic capture control by continuously detecting face regions in real-time and adapting capture decisions based on current detection results. Rather than using fixed timer intervals, the system dynamically adjusts capture timing and positioning according to detected face regions, allowing flexible device adjustment without sacrificing timing control
Data Source
AI summary
An apparatus for generating a 3-dimensional face model includes a multi-view image capturer configured to sense a motion of the mobile device and automatically capture still images from two or more directions; and a 3D model generator configured to generate a 3D face mode using the two or more still images obtained by the multi-view image capturer.