KR100434907B1 - Monitoring system including function of figure acknowledgement and method using this system - Google Patents

Abstract

The present invention provides a video surveillance system having a face recognition function and a method of using the system, by comparing the face data registered by the video registration unit and stored in a database with the face data generated and tagged by the video monitoring unit by the image recognition unit. The present invention provides a video surveillance system capable of face recognition and a method of using the same.

The present invention relates to a system for combining a video surveillance function and a face recognition function so as to efficiently recognize a face, and a method of using the same. The present invention includes face information recognizable from an image data frame generated by the video surveillance system. By tagging the data frames in advance, when face information is to be retrieved later by the face recognition system, only the tagged frames can be searched instead of all data frames. To provide a system and a method of using the same.

Description

MONITORING SYSTEM INCLUDING FUNCTION OF FIGURE ACKNOWLEDGEMENT AND METHOD USING THIS SYSTEM}

The present invention relates to a video surveillance system including a face recognition function and a method of using the same. More particularly, the present invention relates to a video surveillance system including a face recognition function in a video data captured by a surveillance camera. By doing so, when the face shown in the image data frames captured by the surveillance camera to be compared with the stored face data, only the tagged screen can be compared, thereby making it possible to quickly compare the faces. It relates to a system using the same.

In the related art, the face recognition system and the video surveillance system are separated from each other, and a large number of manual operations are involved in determining whether a person to be monitored exists in the video data generated as a result of video surveillance. Was taken.

That is, in order to store image data input through a camera on a recording tape or disk in a video surveillance system, and to read whether the person on the data is a person to be monitored, a person had to visually compare the data. Alternatively, by extracting face information from the stored data and building a new database, and combining this database with a face recognition system, the presence of a person to be monitored could be read.

However, since the amount of data stored by the video surveillance system is huge, it is time-consuming to compare face data contained in these data with face data stored in the database one by one, and every time new image data is generated. Extracting and recognizing the face information recognizable from the data increases not only time but also required storage space day by day, which increases the cost required for this.

Accordingly, an object of the present invention is to tag in advance a data frame including recognizable face information among image data generated by the video surveillance system, thereby determining whether a person to be monitored exists among the image data generated by the video surveillance system. The present invention provides a system and a method of using the same to search for a face at low cost by searching only a tagged frame when searching for.

1 is a block diagram schematically showing one embodiment of a video surveillance system according to the present invention;

FIG. 2 is a block diagram illustrating in more detail an image register and an image recognizer illustrated in FIG. 1; FIG.

3 is a block diagram illustrating in more detail the video monitoring unit illustrated in FIG. 1.

4 is a flowchart illustrating an embodiment of a video surveillance method using a video surveillance system in accordance with the present invention.

The present invention to achieve the above object is a database for storing face image data; An image register for registering face image data in the database; Surveillance camera, a frame acquisition device for generating an image data frame using the image data input from the surveillance camera, the face image detection to detect a frame containing the recognized face information of the image data frame and tag the detected frame A video monitoring unit including a device, a device storing the image data frame, and a device transmitting or outputting the same to a remote location; And an image recognition unit receiving face image data from the image monitoring unit, and comparing the face image data with the face image data stored in the database to recognize a face.

Hereinafter, the configuration and operation of the video surveillance system according to the present invention will be described with reference to FIGS. 1 to 3.

1 is a block diagram showing an embodiment of a video surveillance system according to the present invention, Figure 2 is a block diagram showing in more detail the image register and image recognition unit shown in Figure 1, Figure 3 is shown in FIG. Referring to the configuration and operation of the video surveillance system according to the present invention with reference to these drawings as a block diagram showing the video surveillance in more detail, the system is a database 400 for storing the face image data, video surveillance unit 500, a face image data extracted from a tagged image data frame input from an image registerer 200 and a image monitor 500 that register face image data, and face image data stored in the database 400. In comparison, the image recognition unit 300 recognizes a face.

Looking at the configuration of the video surveillance system according to the present invention in more detail, the image registration unit 200 is the image input unit 202, the face region extraction unit 204, pre-processing unit 206, feature extraction unit 208, face registration unit The image recognition unit 300 includes an image input unit 302, a face region extraction unit 304, a preprocessor 306, a feature extraction unit 308, and a face recognition unit 310. .

Referring to the operation of the image register unit 200, the image data is input from the image input unit 202, the face region extraction unit 204 extracts only the face region from the input image data, the pre-processing unit 206 is the extraction Normalized information such as contrast and distance of the acquired face region image data is normalized. The feature extractor 208 extracts a feature of a face from the preprocessed face region image data. The face register 210 stores the feature of the face extracted from the feature extractor 204 in the database 400.

The image registration unit 200 extracts feature information of the face from the face region image data, stores the feature information and face region image data in the database 400, and then inputs the face input from the image monitoring unit 500. Compare with image data.

On the other hand, the image recognition unit 300 has a configuration substantially the same as the configuration of the image registration unit 200, the difference is that it includes a face recognition unit 310 in place of the face registration unit 210. Referring to the operation of the image recognition unit 300, the image input unit 302 receives an image data frame including face information recognizable from the image monitoring unit 500. Next, operations of the face region extractor 304, the preprocessor 306, and the feature extractor 308 are the same as those of the image register 200, and the face recognizer 310 performs the feature extractor 308. The feature of the face extracted from) is compared with the feature of the face of the face image data stored in the database 500 to recognize whether the face is identical.

The image monitoring unit 500 detects a frame including a recognizable face information and a frame obtaining apparatus 504 for generating an image data frame using the surveillance camera 502 and the image data input from the surveillance camera 502. And a detection device 506 for tagging the detected frame, a device 508 for storing the data frame, a device 510 for transmitting the data frame, and a device 512 for outputting the data frame.

In general, a video surveillance system captures an event using a surveillance camera and performs a function of framing and storing or transmitting the captured image data. However, in the present invention, the frame including the recognizable face information is detected from the generated frame and tagged to the detected frame, so that the image data input from the image monitoring unit 500 and the database 400 can be stored. When comparing face information, all video data frames generated by the video monitoring unit 500 need not be compared but only tagged frames to be compared, thereby providing an image system that can greatly reduce the time required for face recognition. do.

The present invention also provides a method of using the system.

The method of using the video surveillance system according to the present invention comprises the steps of: inputting an image by a camera or the like, generating an image data frame from the input image data, and a frame including recognizable face information among the generated image data frames. Tagging, extracting a facial region from a tagged frame, extracting a facial feature from the extracted facial region, comparing the extracted facial feature with features of other faces stored in a database, and outputting a comparison result It provides a method comprising the steps of.

Hereinafter, a video surveillance method according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart illustrating a video surveillance method according to the present invention.

As shown in FIG. 1, the present invention provides a method of inputting an image by a camera or the like (100), generating an image data frame from the input image data (110), and recognizing face information among the generated image data frames. Tagging a frame including the step 120, extracting a face region from the tagged frame 130, extracting a feature of the face from the extracted face region 140, and extracting the feature of the extracted face. And a face recognition function, the method comprising: comparing the facial feature with other facial features stored in the database (150) and outputting the comparison result (160).

Here, the step of inputting an image by a camera or the like (100) and the step of generating an image data frame from the input image data (110) is a general configuration of a video surveillance system, which captures an event with a surveillance camera and frames the data. The steps are separated.

The next step is to tag 120 a frame including recognizable face information in the image data frame. In this step, by extracting and tagging a frame including the recognizable face information among the frames generated in the frame generation step 110, only the tagged frames among the generated image data frames are compared data for face recognition. To be used. As described above, the following steps are performed in the storage device with tag processing for frames including face information among the image data frames and monitoring the image using the stored data.

The next step is to extract 130 the facial region from the tagged frame. In this step, image data such as background and hair are removed in the tagged frame, and only the face region is extracted. The reason for extracting only the face area as described above is that when the face is recognized, other information such as a background head shape is different from the face area so that different recognition results may be generated.

As a method of extracting only the face region, a method of setting a feature point in advance and extracting data of an area spaced apart from the feature point by a predetermined distance or distance ratio is mainly used. At this time, the most commonly used feature is eyes, and since the two parts clearly distinguished from the face color exist by a certain distance, the image of this pattern is used as a feature point for extracting the face, and the distance between these eyes is multiplied by a certain ratio. Extract portions that are spaced from the eye by a certain percentage. For example, if the distance between the two eyes is D, it is possible to extract 0.5 × D outward of the two eyes, 0.75 × D upwards of both eyes, and 1.5 × D downwards of the eyes as the face region.

The next step is after extracting the facial region as described above, and then extracting facial features from the extracted facial region (140). In this step, the separated facial region is divided into a plurality of cells, and then facial features are extracted based on the contrast information of the facial region divided in each cell.

The next step is to compare 140 the extracted facial features with other facial feature information stored in the database. This type of comparison may vary. For example, by comparing a single facial feature information stored in the database with a plurality of facial feature information input from the video surveillance system, whether or not a specific person having the facial feature stored in the database is captured by the video surveillance system, By comparing a plurality of face feature information stored in a database with one face feature information input from the video surveillance system, there may be a method of confirming who is the person captured by the video surveillance system. By applying the latter comparison method, an access control system using a face recognition method can be constructed.

The next step is to output the comparison result 150. This step is to output the result according to whether or not the facial feature information compared with each other as a result of comparison. In this case, when the facial feature information to be compared as a result of the comparison is matched, the image data having the matched facial feature information may be output together so that a human can finally visually compare for a more accurate comparison. In order to obtain the monitoring result for the frames tagged as described above, operations 130 to 160 may be repeatedly performed.

The above description is just one example of implementing the technical idea of the present invention, and those skilled in the art will be able to infer various modifications based on the above contents. Therefore, the scope of the present invention is not limited to the above description, but will be determined by the contents of the claims disclosed later.

The present invention does not simply combine the face recognition system and the video surveillance system as described above, but pre-tags the video data frame including the recognizable face information among the video data frames generated by the video surveillance system or performs such a function. The apparatus further includes an apparatus to perform the operation, thereby reducing the time and cost for face recognition because only the tagged frame is used as the comparison target data for face recognition instead of all the image data generated in the video surveillance system as the comparison target data for face recognition. Can be significantly shortened or reduced.

Claims (6)

In the video surveillance method including a face recognition function, Inputting an image by a camera or the like; Generating an image data frame from the input image data; Tagging a frame including recognizable face information among each generated image data frame; And Storing the image data frame; To determine whether a face is recognized with respect to the stored image data frames a) extracting a face region from the tagged frame; b) extracting facial features from the extracted facial region; c) comparing the extracted facial features with other facial features stored in a database; And and d) outputting a result of the comparison. The method of claim 1, The comparing may be performed by comparing features of faces of various persons input from the camera with features of one person's face stored in the database and searching whether the same person as the person having the facial features is captured by the camera. Characterized in that the method. The method of claim 1, And comparing the face features of one person input from the camera with the face features of various persons stored in the database and searching for who is the person captured by the camera. delete delete A database for storing facial image data, An image register for registering face image data in the database; Surveillance camera, a frame acquisition device for generating an image data frame using the image data input from the surveillance camera, the face image detection to detect a frame containing the recognized face information of the image data frame and tag the detected frame A video monitoring unit including a device, a device for storing the image data frame, a device for outputting the image data frame, and a device for transmitting the image data frame to a remote location; A face recognition unit configured to receive face image data from the image monitoring unit, and compare the face image data with the face image data stored in the database to recognize a face of a person; The image recognizing unit includes an image input unit receiving a tagged image data frame from the image monitoring unit, a face region extracting unit extracting only face region data from the received image data frame, and a dictionary for extracting facial features from the extracted face region data A face recognition function comprising a preprocessor for performing a task, a feature extractor for extracting facial features from preprocessed face region data, and a face recognition unit for comparing the extracted facial features with facial features stored in a database Video surveillance system including.