Projection Steering Using CV Gaze and Pointing Surface Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing projectors lack the ability to dynamically utilize multiple surfaces in a room without manual repositioning or fixed transitions, limiting user interaction and flexibility in projection.

Innovation Solution

A steerable projector system utilizing computer vision (CV) to determine user position, gaze, and pointing gestures, combined with voice commands, to dynamically select and project onto suitable surfaces.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a projector is fixed in position, then it can project onto a stable surface, but it cannot dynamically utilize multiple surfaces in the room

Engineering Contradiction:
Improveability to project onto multiple surfacesVSAvoidsystem complexity for dynamic surface selection
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The projector system automatically detects user position, gaze direction, and room surfaces using computer vision, then autonomously selects the optimal projection surface without manual intervention. The system serves itself by making intelligent decisions about where to project based on real-time environmental understanding.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical repositioning of the projector with computer vision-based automated surface selection. Instead of physically moving the projector to different locations, the system uses cameras and AI to detect user intent and automatically selects appropriate projection surfaces.

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

2Ease of operation

If manual repositioning is used to change projection surfaces, then the projector can adapt to different surfaces, but user interaction is limited and flexibility is reduced

Engineering Contradiction:
Improveuser interaction flexibilityVSAvoidtime for manual repositioning
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The system continuously monitors user position, gaze direction, and body orientation using computer vision, providing real-time feedback to adjust projection surface selection. This creates a closed-loop system where user actions are immediately detected and responded to, enabling natural and flexible interaction.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system pre-identifies multiple candidate projection surfaces in the room and prepares them for selection before user interaction occurs. When the user indicates interest in a particular area, the system can immediately switch to the appropriate pre-identified surface without delay.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If multiple fixed displays are deployed to provide flexibility, then user interaction options increase, but device complexity and cost increase

Engineering Contradiction:
Improveprojection flexibilityVSAvoidnumber of projector devices
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

A single projector device is made multi-functional by enabling it to project onto any suitable surface in the room through automated surface selection. The projector serves multiple purposes and locations sequentially, replacing the need for multiple dedicated display devices.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The projector system transitions from a static, fixed-position device to a dynamic system that can adapt its projection target in real-time. The ability to dynamically select and switch between multiple surfaces provides the flexibility previously requiring multiple fixed displays.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12581044B1Computer vision-based projection steering
Publication Date: 2026.03.17 AMAZON TECH INC
  • US12581044B1 patent drawing
  • US12581044B1 patent drawing
  • US12581044B1 patent drawing

AI summary

Techniques are generally described for steering projection based on computer vision determination of body position, gaze, and pointing gestures. An example method includes identifying a set of candidate projection surfaces and receiving a voice command to begin projection. The example method also includes capturing a wide-area scene image with a top-down perspective and determining an array of location coordinates indicative of a location of a human in the wide-area scene image, where the array of location coordinates is further indicative of a pointing direction of the human in the wide-area scene image. Finally, the example method includes determining a target projection surface from the set of candidate projection surfaces based on proximity to the location of the human and the pointing direction of the human and projecting a projection image onto the target projection surface.