Camera UI Indicator Detection Using Diagonal Search and Cropped Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing user interface devices struggle to accurately detect the location and shape of indicators, such as fingers, in camera images for precise interaction with computer systems.

Innovation Solution

A user interface device and system that utilizes a detector to search camera images in multiple diagonal directions, crops partial images based on detected ends, and employs shape recognition using deep learning to classify and generate information on indicator shapes and states, enabling accurate detection and interaction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional image processing methods are used to detect indicator location and shape, then the detection process is simple, but the detection accuracy is insufficient

Engineering Contradiction:
Improvedetection accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the image processing into distinct functional stages: binarization to separate foreground from background, diagonal direction search to locate fingertips, partial image cropping to focus on relevant regions, and shape classification to identify indicator states. This segmentation allows each stage to optimize for its specific task, improving overall detection accuracy while managing computational complexity through targeted processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces diagonal direction search as an additional dimension to the conventional axis-aligned scanning approach. By searching in multiple diagonal directions simultaneously, the system can more accurately detect fingertip locations and determine indicator shapes, thereby improving measurement precision without proportionally increasing processing complexity.

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

2Productivity

If the system processes entire camera images for indicator detection, then the detection coverage is complete, but the processing time and computational load increase

Engineering Contradiction:
Improveprocessing speedVSAvoiddetection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent extracts and processes only the necessary portions of the camera image through a systematic approach: first binarizing the image to identify relevant regions, then using diagonal search to locate fingertips, and finally cropping only the partial images containing the indicators for detailed shape analysis. This extraction methodology maintains detection accuracy by focusing computational resources on the most informative regions while significantly reducing overall processing time.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary binarization and diagonal direction search before cropping partial images. These preliminary actions identify and isolate the regions of interest in advance, allowing the subsequent cropping and shape classification to proceed more efficiently. This preliminary processing reduces the computational load on later stages while ensuring that no critical detection information is lost.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12625567B2User interface device, user interface system, and recording medium
Publication Date: 2026.05.12 DEEP INSIGHT INC
  • US12625567B2 patent drawing
  • US12625567B2 patent drawing
  • US12625567B2 patent drawing

AI summary

A location of an indicator in a camera image and a shape of the indicator are accurately detected.A user interface device configured to specify an indication by identifying a shape of an indicator based on a camera image obtained by capturing the indicator having the shape that changes for the indication includes a detector configured to search an image obtained by binarizing the camera image in at least two different diagonal directions to detect an end of an object included in the camera image, an image cropping unit configured to crop a partial image from the camera image based on the detected end, a selector configured to select a partial image indicating a shape before the indication and a partial image indicating a shape of the indication by the indicator when the partial image is a partial image obtained by capturing the indicator as the object, and an information generator configured to generate information indicating whether the indicator has the shape before the indication or the shape of the indication.