Camera-Based Virtual Keyboard Input with Finger Position Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual keyboards on electronic devices suffer from high typographical error rates due to uniform detection of hand movements without considering device posture and finger position and bending, making it difficult to distinguish between keys in the same column and adjust the keyboard layout.
Innovation Solution
An electronic device with a camera and processor that crops images to detect finger objects, matches key columns to image regions, identifies key rows based on finger positions and bending, and determines virtual key inputs through typing movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If uniform detection of hand movement is used without considering device posture and finger position, then the detection process is simple, but the typographical error rate increases
Solution Approach 1:
The system changes detection parameters dynamically based on device posture (horizontal/vertical) and finger characteristics (position, bending degree). Instead of using fixed detection parameters, the system adjusts key column and key row determination parameters according to the detected device orientation and finger state, thereby reducing typographical errors without significantly increasing system complexity
Solution Approach 2:
The detection system transitions from a static uniform detection approach to a dynamic adaptive detection approach. The system continuously adapts detection parameters based on real-time device posture and finger position, making the detection process flexible and context-aware while maintaining reasonable complexity
2Stability of the object's composition
If key columns are fixedly matched without considering finger position and bending, then the keyboard layout is stable, but the ability to distinguish keys in the same column is reduced
Solution Approach 1:
The system applies different matching criteria for different regions of the keyboard. Instead of using a single fixed matching rule for all keys, the system determines key columns and key rows using different parameters based on local finger position and bending degree, enabling precise differentiation of keys within the same column while maintaining overall layout stability
Solution Approach 2:
The keyboard detection is segmented into two independent dimensions: key column determination (based on horizontal finger position) and key row determination (based on finger bending degree). This segmentation allows the system to independently optimize each dimension's detection parameters, improving key identification accuracy without compromising layout stability
3Device complexity
If device posture is not considered in image processing, then the processing is simpler, but the accuracy of finger position detection is reduced
Solution Approach 1:
The image processing system dynamically adjusts the cropping region based on detected device posture. When the device is in horizontal or vertical orientation, the system automatically adjusts the image cropping parameters to capture the relevant finger region accurately, thereby improving detection precision without requiring complex processing for all possible orientations
Data Source
AI summary
Disclosed is an electronic device configured to obtain a second image, by cropping a first image obtained through a camera, detect at least one finger object in the second image, the at least one finger object corresponding to at least one finger, determine at least one key column corresponding to the at least one finger object among a plurality of key columns of a virtual keyboard by matching the plurality of key columns of the virtual keyboard to a plurality of regions of the second image, and identifying at least one region corresponding to at least one position of the at least one finger object among the plurality of regions of the second image, determine at least one key row corresponding to the at least one finger object among a plurality of key rows of the virtual keyboard, identify, based on the determined at least one key column and the determined at least one key row, at least one virtual key corresponding to the at least one finger object among a plurality of virtual keys of the virtual keyboard, and identify an input of the identified at least one virtual key, based on detecting a typing movement of the at least one finger object.


