CN108335374A - A kind of automatic roll-calling method - Google Patents

Abstract

The present invention relates to the field of educational procedural monitoring, in particular to an automatic roll call method; comprising the following steps: acquiring the facial feature information of a person entering the classroom through a face detection model; acquiring the facial feature information of the person through a facial feature extraction model ; Query the corresponding student information according to the acquired facial feature information; the successfully identified students will automatically sign in in the system. The present invention automatically finds relevant information about students by automatically identifying the facial feature information of the attendees, counts the students attending the class, solves the problems answered during roll call, and achieves the effect of automatic roll call.

Description

An automatic roll call method

technical field

The invention relates to the field of educational process monitoring, in particular to an automatic roll call method.

Background technique

In daily teaching, in order to count the attendance of students, the teacher will roll the students in the class to count the attendance and absence of the class. This method generally conducts the roll call by directly confirming the number of students by roll call, especially when the number of students is large, it will waste more teaching time, the efficiency is low, and there is also the situation of substitute roll call. However, due to the large number of students who use attendance machines to check in, some students may punch in and leave without actually attending class, and attendance is in name only.

Contents of the invention

The purpose of the present invention is to provide an automatic roll call method. The present invention automatically finds the relevant information of students by automatically identifying the facial feature information of the attendees, counts the students who attend the class, and solves the problems answered during the roll call, so as to achieve automatic roll call. Effect.

In order to achieve the above object, the present invention adopts the following technical solutions.

An automatic roll call method, comprising the following steps;

Obtain the face feature information of the people entering the classroom through the face detection model;

Obtain the facial feature information of the person through the facial feature extraction model;

Query the corresponding student information according to the acquired facial feature information;

Successfully identified students are automatically signed in to the system;

Further, the face feature information of the people entering the classroom is obtained through the face detection model. Before that, it also includes training the face detection model based on the deep learning framework and network architecture.

Further, the facial feature information of the person is obtained through the facial feature extraction model. Before that, it also includes training the facial feature extraction model based on the deep learning framework and network architecture.

Further, according to the obtained facial feature information, the corresponding student information is queried. Before that, it also includes establishing a student facial feature information database.

Further, the establishment of a student facial feature information database, and later, the establishment of a subject-corresponding student information database.

Further, the successfully identified students are automatically signed in in the system, and then, the corresponding number of lectures attended by the student is automatically incremented by 1.

Further, querying corresponding student information according to the acquired facial feature information includes first querying corresponding student information from the subject-corresponding student information database; and then querying corresponding student information from the student facial feature information database for unidentified students.

Further, if the student’s facial feature information database is not identified, the reason for the failure is analyzed; if the facial feature information is not clear, the teacher is notified; if the facial feature information is clear, it is entered into the unknown person information database.

Further, the teacher receives the unidentified information, and by asking questions, etc., makes the student's facial features more prominent and re-identifies.

Further, face recognition is performed from one or more than one direction respectively; and the recognition results of each time are compared.

Compared with the prior art, the present invention has the following advantages:

1. Automatically recognize the facial features of students in the classroom, and automatically roll their names, saving time and improving efficiency;

2. Automatically count the number of attendance of the student in this subject, and the teacher will refer to the student's usual performance evaluation;

3. Multiple identifications to increase the success and accuracy rate of identification.

Description of drawings

Fig. 1 is a schematic flow sheet 1 of a kind of automatic roll call method of the present invention;

Fig. 2 is a schematic flow sheet 2 of a kind of automatic roll call method of the present invention;

Detailed ways

The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

An automatic roll call method, comprising the following steps:

Example 1

Please refer to Fig. 1, Fig. 1 is the schematic flow sheet of a kind of automatic name-calling method of the present invention; The present embodiment provides a kind of automatic name-calling method, comprises the following steps:

Step S101, obtain the face feature information of the personnel entering the classroom through the face detection model;

Step S102, obtaining facial feature information of the person through a facial feature extraction model;

Step S103, query corresponding student information according to the acquired facial feature information;

Step S104, students who have been successfully identified are automatically signed in in the system.

Example 2

Please refer to Fig. 2, and Fig. 2 is the schematic flow chart 2 that is a kind of automatic roll-calling method of the present invention; Present embodiment provides another kind of automatic roll-calling method, comprises the following steps:

Step S201, based on the deep learning framework and network architecture, the face detection model is obtained through training;

The face detection model adopts the caffe deep learning framework and the Faster-RCNN network architecture, but is not limited to using this deep learning method and this network architecture; like Caffe at the University of California, Berkeley, Theano at the Montreal Institute of Technology, and IDSIA, the Swiss artificial intelligence laboratory Brainstorm, Princeton University Marvin, etc. can be used as a framework for this deep learning; SSD, Faster-RCNN, etc. can be used as a network architecture for this modeling.

Step S202, based on the deep learning framework and network architecture, the facial feature extraction model is obtained through training;

The face feature extraction model adopts the caffe-based deep learning framework and the VGGFACE network architecture, but is not limited to the use of this deep learning method and this network architecture;

Extract multiple sets of face pictures in different scenes, perform in-depth learning, and output each facial feature point; set various facial feature points, including the dynamic and static facial feature information of the person's face, including the shape, size, skin, and hair color of the face, The center point information and boundary point information of the face and the boundary point information of the face also include local feature information, such as nose, mouth, left eye, right eye, left eyebrow and right eyebrow and mutual distance.

Step S203, establishing a student facial feature information database;

From the student status information database, according to the face detection model and facial feature extraction model, extract the facial feature information of the student, and associate with the student information to establish the student facial feature information database;

The student facial feature information database includes student name, student registration number, and facial feature information.

Step S204, establishing a subject-corresponding student information database;

According to the educational information and student information, establish a subject-corresponding student information database. The subject-corresponding student information database includes student registration number, subject information, and facial feature information.

Step S205, obtain the face feature information of the personnel entering the classroom through the face detection model;

Obtain the face feature information of the people entering the classroom through the face detection model; after students enter the classroom, the door recognition module first recognizes the face frame; after entering the classroom, the top recognition module recognizes the face frame; the front and rear doors are equipped with recognition modules .

Step S206, obtaining the facial feature information of the person through the facial feature extraction model;

After the recognition module recognizes the face frame diagram, the facial feature information of the person is obtained through the facial feature extraction model; multiple recognition sources recognize multiple sets of facial feature information, compare multiple sets of facial feature information, and combine them into the most suitable facial features information.

Step S207, querying the corresponding student information from the subject corresponding student information database;

According to the obtained facial feature information, compare this information with the facial feature information in the corresponding student information database of the subject, and obtain the student registration number and subject information of the corresponding student;

Step S208, judging whether the student information is found;

If the student information is found, then execute step S209, step S210, and this inquiry ends; otherwise, enter step S211;

Step S209, students who are successfully identified are automatically signed in in the system;

From the identification to the student, according to the identified student's student status number and subject information, automatic sign-in is performed in the system;

Step S210, the number of corresponding lectures for the student is automatically increased by 1;

If the student is successfully identified from the student information database corresponding to the subject of the student, the number of lectures in the corresponding subject will be automatically accumulated according to the identified student status number and subject information;

Step S211, querying the corresponding student information from the student facial feature information database;

If the student is not identified in the corresponding student information database of the subject, the obtained facial feature information will be compared with the facial feature information in the student facial feature information database to obtain the student status number and subject information of the corresponding student;

Step S212, judging whether the student information is found;

If the student information is found, execute step S209 and step S210, and this inquiry ends; otherwise, enter step S213.

Step S213, analyzing the unrecognized cause;

If the student has not been identified from the student's facial feature information database, then reason analysis is required; it is divided into two types: unclear and clear.

Step S214, judging whether the facial feature information is unclear;

If the student's facial feature information is not clear, then execute step S215, step S216, if the student's facial feature information is clear, then enter step S217;

Step S215, notify the teacher of unidentified student information;

If the student's facial feature information is clear, the teacher will be notified of the unrecognized student information; the notification information includes the reason for not being recognized, the seat number, etc.;

Step S216, the teacher makes the student's facial features more prominent by asking questions

After receiving the unidentified information, the teacher makes the student's facial features more prominent by asking questions, etc., and proceeds to step S205 for re-identification;

Step S217, the facial feature information of the unidentified personnel is clear, and entered into the unknown personnel information database;

For the facial feature information of the person whose facial feature information is clear but has not yet been found, enter the unknown person information database as an unidentifiable person.

For multiple identifications, the facial feature information of persons who still fail to be identified is entered into the facial feature fuzzy information database.

What is shown above is only a preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, some improvements and modifications can also be made without departing from the principles of the present invention. It should be regarded as the protection scope of the present invention.

Claims (10)

1. a kind of automatic roll-calling method, which is characterized in that include the following steps；

A. the face characteristic information into this classroom personnel is obtained by Face datection model；

B. the face feature information of the people is obtained by facial feature extraction model；

C. corresponding student information is inquired according to the face feature information of acquisition；

D. identify that successful student is automatically signing in systems.

2. a kind of automatic roll-calling method according to claim 1, which is characterized in that obtained and entered by Face datection model The face characteristic information of this classroom personnel further includes being based on deep learning frame and the network architecture before, and training obtains face Detection model.

3. a kind of automatic roll-calling method according to claim 1, which is characterized in that obtained by facial feature extraction model The face feature information of the people further includes being based on deep learning frame and the network architecture before, and training obtains facial characteristics and carries Modulus type.

4. a kind of automatic roll-calling method according to claim 1, which is characterized in that looked into according to the face feature information of acquisition It further includes establishing student's face feature information library before to ask corresponding student information.

5. a kind of automatic roll-calling method according to claim 4, which is characterized in that student's face feature information library is established, Later, further include establishing subject to correspond to student information library.

6. a kind of automatic roll-calling method according to claim 1, which is characterized in that identify successful student in systems certainly It is dynamic to register, later, further include that the student corresponds to number of listening to the teacher and adds 1 automatically.

7. a kind of automatic roll-calling method according to claim 5, which is characterized in that looked into according to the face feature information of acquisition Corresponding student information is ask, including, it is corresponded to from subject inquire corresponding student information in student information library first；To unidentified arrive It is raw to inquire corresponding student information from student's face feature information library again.

8. a kind of automatic roll-calling method according to claim 7, which is characterized in that student's face feature information library is unidentified The student arrived analyzes unidentified reason；It is unintelligible if it is face feature information, then notify teacher；If facial special Reference breath is clear, enters unknown personnel's information bank.

9. a kind of automatic roll-calling method according to claim 8, which is characterized in that teacher receives unidentified information, passes through The modes such as enquirement, make student's facial characteristics more prominent, are again identified that.

10. a kind of automatic roll-calling method according to claim 1, which is characterized in that from one and 1 when progress recognition of face A above direction is identified respectively；And each secondary recognition result is compared.