CN113157196A

CN113157196A - Finger recognition algorithm

Info

Publication number: CN113157196A
Application number: CN202110454741.7A
Authority: CN
Inventors: 刘传义; 李雅洁
Original assignee: Lanzhou University
Current assignee: Lanzhou University
Priority date: 2021-04-26
Filing date: 2021-04-26
Publication date: 2021-07-23

Abstract

The invention discloses a finger recognition algorithm, relates to a human-computer interaction interface, and particularly relates to judgment of abundant input primitives in the human-computer interaction interface and an operating hand in a multi-point touch interface and recognition of fingers, so as to solve the problems that the existing recognition algorithm depends on specific equipment, has high operation complexity, poor real-time performance and the like. The algorithm identifies fingers corresponding to the contact points of the fingers in the multi-point touch interface through information of natural gestures such as relative orientation relation and distance between the finger contact points, and identifies whether the operating hand is a left hand or a right hand according to the positive and negative of the cross product of vectors between the thumb contact point and the index finger contact point and the little finger contact point respectively. The invention can enrich the input primitive in the multi-point touch interface, provide convenience for people to realize 'blind operation' man-machine interaction such as touch text input in the multi-point touch interface, effectively utilize multi-point touch equipment for the blind and the like, and improve the interaction performance.

Description

Finger recognition algorithm

Technical Field

The invention relates to a human-computer interaction interface, in particular to rich input primitives in the human-computer interaction interface and recognition of an operating hand and fingers in a multi-point touch interface.

Background

Multi-touch screen devices such as smart phones and tablet computers are widely used in daily work, study and life of people. A touch screen is both an output device and an input device. Touch screen equipment does not have external input devices such as mouse, keyboard usually, and people adopt bare-handed operation above when using touch screen equipment usually: this provides on the one hand convenience of operation; but on the other hand, the input primitives are not rich enough, so that the interaction performance of people is limited. For example, when people input text on a mobile phone or a tablet computer, the existing method mainly uses a single finger to click a soft keyboard for inputting, and the input mode is generally very inefficient compared with the mode of inputting by two hands through a physical keyboard.

If the fingers touching the screen can be recognized and it is determined whether the operator's hand is left or right, then the characters on the physical keyboard can be assigned to the corresponding fingers according to the correct fingering entered by the person on a standard keyboard (QWERTY keyboard): therefore, people can realize an input mode similar to that on a physical keyboard on the multi-point touch screen, the input efficiency of the multi-point touch screen can be close to that on the physical keyboard, and touch typing input can be carried out similarly. More input primitives can be provided for the multi-touch screen through finger recognition, so that people can obtain more interactive instructions during free-hand operation. The finger recognition algorithm supports people to perform blind operation on the multi-point touch screen, and great convenience is provided for the blind to use the multi-point touch screen.

Typical identification methods currently available require specially-made equipment (e.g., documents "Masson D, Goguey A, Malaria S, et al. Whichfings: Identifying transistors on Touch Surfaces and keys Using vibrations [ C ]// Proceedings of the 30th Annual ACM Symposium on User Interface Software and technology. Quebec City, QC, Canada: ACM,2017: 41-48."), or a camera specifically positioned relative to the operative surface (e.g., documents "Zheng J, volume D. Finger-Aware Shortcuts [ C ]// Proceedings of the CHI 2016 family J, CA, san: ACM,2016: 4285): such a condition is not provided in a general touch device. The finger recognition algorithm based on image processing has high operation complexity, and can realize good real-time performance of finger recognition only by equipment with high processing capacity.

Disclosure of Invention

The invention provides a finger recognition algorithm based on the contact points of fingers on a touch screen or other operation surfaces, which comprises the steps of recognizing a thumb contact point according to the sum of the distances from one contact point to the other 4 contact points; identifying a forefinger contact and a little finger contact according to the distances from the rest contacts to the thumb contact; according to the remaining 2 unidentified contacts, the distances from the projection points on the connecting line between the contact points of the thumb and the little finger to the contact points of the thumb can be sequentially identified as the contact points of the middle finger and the ring finger from small to large; the left hand or the right hand of the operating hand can be judged according to the positive and negative of the cross product of the two vectors with the thumb contact as the starting point and the forefinger contact and the little finger contact as the end points.

And judging that the contact point of the thumb is the largest sum of the distances from a certain finger contact point to the contact points of the rest four fingers, the contact point of the index finger, which is closest to the contact point of the thumb, is the contact point of the little finger, which is farthest from the contact point of the thumb.

The finger recognition algorithm provided by the invention has low operation complexity, and can realize real-time calculation on the conventional common touch equipment (through testing, the recognition can be completed within 1 millisecond).

In the experiment participated in by 20 testees of different ages and sexes, under the condition that the multi-point touch screen is placed towards the testees at any angle, each tester touches the screen 1000 times by 5 naturally-opened gestures, and the accuracy rate of the algorithm for judging the operating hand and identifying the fingers is 100 percent; in the process of simulating the text input of a physical keyboard on a touch screen by each testee by using a standard fingering method, 1000 gestures are respectively and continuously recorded, the judgment accuracy of the statistical result display algorithm on an operating hand is 100%, and the accuracy of finger identification is 96.8%. Here, correct finger recognition means that 10 fingers are recognized correctly in both hands at the same time.

Drawings

FIG. 1 is a diagram of the effect of algorithm execution, wherein fingers and contact points thereof are identified by the initial letters of the fingers in English (i.e. thumb-T, index finger-I, middle finger-M, ring finger-R, little finger-L, and the fingers are identified in the same way in FIGS. 2 and 3);

FIG. 2 is a schematic diagram of a middle finger and ring finger identification method;

fig. 3 is a schematic diagram of a left-right hand recognition method (taking a right hand as an example).

Detailed Description

5 fingers of an operator touch a touch screen or other operation surfaces simultaneously to trigger a finger recognition algorithm to be executed, firstly, coordinates of 5 touch points of the fingers on an interface are obtained, then, the sum of distances from any touch point to the rest 4 touch points is calculated, and the largest touch point is judged as a thumb touch point T.

And calculating the distances from the rest 4 contact points to T, wherein the minimum value is I, and the maximum value is L.

Calculating projection points of the rest 2 finger contacts to the line segment TL, recording the projection point with the closer distance to the thumb contact T as M', and judging the corresponding contact as a middle finger contact M; the other projection point is denoted as R', and the corresponding contact point is the ring finger contact point R, see FIG. 2.

Computing

The operator is determined to be the left hand or the right hand respectively according to the positive and negative of the vector cross product, see fig. 3.

Claims

1. A finger recognition algorithm comprises a thumb recognition method, an index finger recognition method, a little finger recognition method, a middle finger and ring finger recognition method and a left-right hand recognition method, and is characterized in that a thumb touch point is recognized according to the sum of the distances from one touch point to the other 4 touch points; identifying a forefinger contact and a little finger contact according to the distances from the rest contacts to the thumb contact; according to the projection points of the rest 2 unidentified contact points on the connecting line between the contact points of the thumb and the little finger, the contact points of the middle finger and the ring finger can be sequentially identified from small to large in the distance from the contact point of the thumb; the left hand or the right hand of the operating hand can be judged according to the positive and negative of the cross product of the two vectors with the thumb contact as the starting point and the forefinger contact and the little finger contact as the end points.

2. The thumb recognition method, the second thumb recognition method and the little finger recognition method according to claim 1, wherein the sum of the distances from one touch point to the remaining 4 touch points is the largest thumb touch point; the closest contact to the thumb contact is the two-thumb contact, and the farthest contact to the thumb is the little finger contact.