AR Text Clarity via Virtual Surface Overlay

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Head-mounted displays (HMDs) face challenges in providing clear pass-through views, particularly with text clarity, which affects user safety and usage time due to the unclear details in real-space images captured by their cameras.

Innovation Solution

An image displaying method that captures real-space images, detects text regions, recognizes text content, obtains feature points, and creates a virtual surface for displaying text content, enhancing text clarity through feature extraction and virtual image generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the camera directly captures real space images for pass through views, then the device complexity is low and operation is simple, but the text clarity and detail visibility deteriorate

Engineering Contradiction:
Improvetext clarityVSAvoidimage processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the real space image into multiple regions including text regions and non-text regions. By detecting text regions specifically and processing them separately through OCR and virtual surface generation, the system enhances text clarity without unnecessarily processing the entire image, thus managing complexity effectively.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediate processing steps including text region detection, OCR recognition, and virtual surface creation as mediators between the original captured image and the final displayed pass through view. These intermediaries transform unclear text into clear virtual text overlays while preserving the original background image.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Loss of information

If the camera directly captures real space images for pass through views, then the processing time is short and response is fast, but the text readability and user safety deteriorate

Engineering Contradiction:
Improvetext information clarityVSAvoidimage processing time
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The patent extracts only the text regions from the complete real space image for enhanced processing. By identifying and isolating text areas through text region detection, the system applies OCR and virtual surface generation only to these extracted text portions, minimizing processing time while maximizing text information clarity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary text region detection and OCR recognition on captured images before displaying the pass through view. By pre-processing and identifying text regions in advance, the system prepares clear text overlays that can be quickly composited onto the final display, reducing overall processing time while ensuring text readability for user safety.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20240242446A1Image displaying method, electronic device, and non-transitory computer readable storage medium
Publication Date: 2024.07.18 HTC CORP
  • US20240242446A1 patent drawing
  • US20240242446A1 patent drawing
  • US20240242446A1 patent drawing

AI summary

An image displaying method is disclosed. The image displaying method includes the following operations: capturing a first image of a real space by a camera based on a first viewing direction when the camera is located at a first camera position, wherein the first image includes a first text image; detecting a text region according to the first image by a processor, wherein the text region includes the first text image; recognizing the first text image to obtain a first text content by the processor; obtaining several first feature points of the text region according to the first image by the processor; creating a first virtual surface according to the several first feature points by the processor; and displaying a first virtual image with the first text content appending to the first virtual surface by a display circuit.