Neural Network Face Centering with Audio Guidance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users with visual impairments cannot adjust camera position to center a subject within the camera's field of view, as existing solutions fail to provide feedback when the subject's face is not visible or out of the frame.

Innovation Solution

A neural network system trained with a dataset of images to provide directional assistance by classifying the position of a face within the camera's field of view, generating prompts to guide users in adjusting the camera or subject position to center the face.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If visual display screen is used to show camera feed, then user can see whether subject is centered, but users with visual impairments cannot use the display screen to see what is being captured

Engineering Contradiction:
Improveease of use for visually impaired usersVSAvoidloss of visual feedback information
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The patent introduces an intermediary system that translates visual information from the camera feed into non-visual feedback formats. The neural network acts as a mediator between the camera input and the user, converting spatial position data into directional audio cues that guide users with visual impairments to center the subject in the frame.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the visual display mechanism with an auditory feedback system. Instead of relying on the mechanical/optical display screen that visually impaired users cannot access, the system substitutes a sound-based interface that provides equivalent functional information through directional audio prompts.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If existing facial analysis solutions are used, then face centering can be detected, but these solutions fail to provide feedback when the subject's face is not within the field of view

Engineering Contradiction:
Improveprecision of face position detectionVSAvoidreliability when face is out of view
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent performs preliminary detection of the subject's presence and approximate location before attempting detailed facial analysis. The system first determines whether a subject is within the field of view using broader detection methods, then activates more precise facial feature analysis only when appropriate conditions are met, providing fallback feedback when the face is completely out of view.

Inventive Principle:
Principle #10Preliminary action

3Ease of operation

If display screen feedback is provided, then user can adjust camera position, but users with visual impairments receive no guidance on how to adjust

Engineering Contradiction:
Improveease of camera positioningVSAvoidloss of directional guidance information
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The patent implements a closed-loop feedback system that provides continuous directional guidance to users. The neural network analyzes the subject's position and generates real-time audio feedback indicating the direction and magnitude of adjustment needed, creating an iterative process where users receive ongoing guidance until the subject is properly centered.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS11074430B2Directional assistance for centering a face in a camera field of view
Publication Date: 2021.07.27 ADOBE INC
  • US11074430B2 patent drawing
  • US11074430B2 patent drawing
  • US11074430B2 patent drawing

AI summary

Methods and systems are provided for providing directional assistance to guide a user to position a camera for centering a person's face within the camera's field of view. A neural network system is trained to determine the position of the user's face relative to the center of the field of view as captured by an input image. The neural network system is trained using training input images that are generated by cropping different regions of initial training images. Each initial image is used to create a plurality of different training input images, and directional assistance labels used to train the network may be assigned to each training input image based on how the image is cropped. Once trained, the neural network system determines a position of the user's face, and automatically provides a non-visual prompt indicating how to center the face within the field of view.