WO2018053763A1

WO2018053763A1 - Image identification method and device

Info

Publication number: WO2018053763A1
Application number: PCT/CN2016/099757
Authority: WO
Inventors: 张勇; 刘磊; 陈泽虹; 赵东宁; 陈剑勇; 李岩山
Original assignee: 深圳大学
Priority date: 2016-09-22
Filing date: 2016-09-22
Publication date: 2018-03-29

Abstract

An image identification method, comprising: acquiring a pixel of an image to be identified; if a pixel value of the pixel is less than or equal to pixel values of multiple pixels in a predetermined region, then determining that the pixel is a target pixel, wherein the predetermined region is a region having the pixel as a center point; and forming, by means of the determined target pixel, an image of a head region in the image to be identified.

Description

Image recognition method and device

Technical field

The invention belongs to the field of image recognition, and in particular relates to an image recognition method and device.

Background technique

With the increasing urbanization, the urban population continues to expand, the activities of the group are increasing day by day, and the safety of the population has become a social problem, so the number of people has become a research hotspot.

Usually, the number of people will use the head recognition technology to count the head after the identification. The existing head recognition technology uses a color model to identify the head. In the preliminary preparation work, a large number of head samples need to be collected for learning, such as wearing, dressing, hairstyle, hat, and head ornament type.

The existing head recognition technology uses a color model to recognize the head image, such as colors that appear in the head or different dresses and hairstyles, which may lead to misjudgment and low accuracy of image recognition.

technical problem

The invention provides an image recognition method and device, aiming at solving the problem of low image recognition accuracy.

Technical solution

In order to solve the above technical problem, the present invention is implemented as follows. An image recognition method includes: acquiring pixel points in an image to be recognized, when a pixel value of the pixel point is less than or equal to a plurality of pixel points in a preset area. Determining, in the pixel value, the pixel point as a target pixel point, wherein the preset area is an area centered on the pixel point, and the determined target pixel point constitutes a head in the image to be recognized The image of the area.

An image recognition apparatus includes: an acquisition module, a determination module, and a constituent module;

Obtaining, by the module, the pixel in the image to be identified, when the pixel value of the pixel is less than or equal to the pixel value of the plurality of pixels in the preset area, the determining module determines that the pixel is the target pixel, where The preset area is an area centered on the pixel, and the constituent module forms an image of the head area in the image to be identified by the determined target pixel.

Beneficial effect

Compared with the prior art, the present invention has the beneficial effects that: the present invention obtains a pixel point in an image to be identified, and when the pixel value of the pixel point is less than or equal to a pixel value of a plurality of pixel points in the preset area, determining the pixel The point is a target pixel, wherein the preset area is an area centered on the pixel point, and the determined target pixel point constitutes an image of the head area in the image to be recognized. In this way, the pixel value of the single pixel point can be compared with the pixel value of the plurality of pixel points to determine the pixel area of the head region, which is not affected by the clothing, hairstyle and color of the pedestrian, and improves the image recognition of the head region. Precision.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is some embodiments of the invention.

1 is a schematic flowchart of an implementation of an image recognition method according to a first embodiment of the present invention;

2 is a schematic flowchart showing an implementation of an image recognition method according to a second embodiment of the present invention;

3 is an image to be identified according to a second embodiment of the present invention;

4 is a pixel value change curve in an image to be recognized according to a second embodiment of the present invention;

FIG. 5 is a schematic diagram of an image recognition apparatus according to a third embodiment of the present invention; FIG.

FIG. 6 is a schematic diagram of an image recognition method according to a fourth embodiment of the present invention.

Embodiments of the invention

The technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. The embodiments are merely a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by a person skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

The head region image recognition method provided by the embodiment of the present invention can be applied to all terminals having a display function such as a camera, a television, a display imaging device, and the like.

Please refer to FIG. 1. FIG. 1 is a schematic flowchart of an image recognition method according to a first embodiment of the present invention, which can be applied to all display image devices having a display function. The image processing method shown in FIG. 1 mainly includes the following steps:

S101. Acquire a pixel in an image to be identified.

The image to be recognized is an image for recognizing a head region captured by a depth camera or a stereo camera, and the image may be a certain frame image in the video captured by the depth camera.

S102. When a pixel value of the pixel is less than or equal to a pixel value of a plurality of pixel points in the preset area, determining the pixel point as the target pixel point;

The target pixel is a pixel in the head region. The preset area is an area centered on the pixel, and the preset area may be a regular pattern such as a circle, a square or an ellipse, or may be a circular circumference, a square side length or an elliptical circumference. In an image taken with a depth camera or a somatosensory camera, the pixel value of a pixel of an object is proportional to the distance from the camera to the object. Therefore, the closer the object in the image to be recognized is to the camera, the lower the pixel value of the object. Since the head area is closest to the camera when standing, the pixel value of the head area pixel is the smallest in the image to be recognized, that is, the pixel value of the pixel area of the head area is smaller than the shoulder and the ground waiting Identifies the pixel values of other areas displayed in the image.

S103. Form, by the determined target pixel, an image of a head region in the image to be identified.

In the first embodiment of the present invention, the pixel point in the image to be identified is obtained. When the pixel value of the pixel point is less than or equal to the pixel value of the plurality of pixel points in the preset area, the pixel point is determined as the target pixel point, where The preset area is an area centered on the pixel point, and the determined target pixel point constitutes an image of the head area in the image to be recognized. In this way, the pixel value of the single pixel point can be compared with the pixel value of the plurality of pixel points to determine the pixel area of the head region, which is not affected by the clothing, hairstyle and color of the pedestrian, and improves the image recognition of the head region. Precision.

As shown in FIG. 2 to FIG. 3, FIG. 2 is a schematic flowchart of an implementation of an image recognition method according to a second embodiment of the present invention, which can be applied to all display image devices having a display function. The image recognition method shown in FIG. 2 mainly includes the following steps:

S201. Acquire pixel points in the image to be identified.

The image to be recognized is an image captured by a depth camera or a somatosensory camera for recognizing a head region, and the image may be a certain frame image in the video captured by the depth camera.

In an image taken with a depth camera or a somatosensory camera, the pixel value of a pixel of an object is proportional to the distance from the camera to the object. Therefore, the closer the object in the image to be recognized is to the camera, the lower the pixel value of the object. Since the head area is closest to the camera when standing, the pixel value of the head area pixel is the smallest in the image to be recognized, that is, the pixel value of the pixel area of the head area is smaller than the shoulder and the ground waiting Identifies the pixel values of other areas displayed in the image.

Specifically, FIG. 3 is an image to be recognized, wherein the pixel points in the line segment AD shown in FIG. 3 include pixel points in the head region image and the shoulder region image, that is, the pixel points in the line segment AB and the line segment CD are shoulder regions. The pixel in the line segment BC is the pixel point in the head region. Fig. 4 shows a pixel value variation curve of pixel points in the line segment AD in the image to be recognized, wherein A', B', C', D' correspond to A, B, C, D, respectively. As can be seen from FIG. 3, the pixel points in the line segment BC are the pixel points in the head region, and the pixel points in the line segment AB and the line segment CD are the shoulder region pixel points. As can be seen from the pixel value variation curve shown in Fig. 4, the pixel values of the pixel points in the line segment B'C' are smaller than the pixel values of the pixel points in the line segment A'B' and the line segment C'D'. Let point K be any pixel point acquired in the image to be recognized.

S202, selecting a preset number of pixel points on each side of the pixel point according to an arrangement order of pixel points in the image to be identified, to form a first pixel point set and a second pixel point set;

Specifically, as shown in FIG. 3, along the order of pixel points in the line segment AD, a preset number of pixel points are respectively selected at both ends of the point K to form a first pixel point set and a second pixel point set. This preset number can be customized. Preferably, the line segment length value of the pixel points in the first pixel point set or the second pixel point set is a value of an adult head radius.

S203. Subtracting between a pixel average value of the pixel points in the first pixel point set and a pixel average value of the pixel point in the second pixel point set to obtain a difference value;

S204. If the absolute value of the difference is less than a preset value, determine that the pixel point is a center point of the preset area;

By judging that the absolute value of the difference is less than the preset value, the position of the pixel point, that is, the position in the line segment AD of the point K, as shown in FIG. 3, can be determined. The magnitude of the absolute value of the difference and the position of the point K are analyzed as follows:

It can be seen from the pixel value variation curve shown in FIG. 4 that when the point K is in the line segment A'B', the smaller the distance between the points K and B' is, the smaller the absolute value of the calculated difference is. Similarly, the point K is online. In the segment C'D', the smaller the distance between the points K and C', the smaller the absolute value of the calculated difference; when the point K is in the line segment B'C', the calculated absolute value is smaller than the point K online. The absolute value calculated for the segment A'B' or the segment C'D'. The preset value can be customized according to the above calculated absolute value, ensuring that point K is in line segment B'C' or near point B' or point C', when point K is in line segment B'C' or in In the vicinity of the point B' or the point C', it is determined that the point K is the center point of the preset area, so that, on the one hand, the calculation amount of the pixel value of the pixel point and the pixel value of the pixel point in the preset area can be reduced; On the other hand, the determination range of the target pixel point can be narrowed, and the recognition efficiency can be improved.

S205 detects a diameter of a head region in the acquired plurality of image samples;

S206. Calculate an average value of the diameter, and use the average value as a preset length;

Specifically, the diameter of the head region in the plurality of image samples is detected by an image processing tool such as AutoCAD or Photoshop.

S207: taking the pixel as a center and determining a circular area by using a preset length as a radius;

S208. The circular area is used as the preset area.

S209, when the pixel value of the pixel is less than or equal to a pixel value of a plurality of pixel points in the preset area, determining the pixel point as the target pixel point;

The target pixel is the pixel of the overhead image. Specifically, when the pixel value of the pixel is less than or equal to the pixel value of the plurality of pixels in the circular area, the pixel is determined to be the target pixel, wherein the number of the plurality of pixels can be customized. Preferably, when the pixel value of the pixel point is less than or equal to the pixel value of the plurality of pixel points in the circumference of the circular area, determining the pixel point as the target pixel point,

Specifically, FIG. 4 is taken as an example. When the point K is in the line segment B′C′, the pixel value of the pixel where the point K is located is less than or equal to the pixel value of the pixel in the circle, and when the point K is When the line segment A'D' or the line segment C'D', the circumference and the line segment B' C' has an intersection point. At this time, the pixel value of the pixel at which the point K is located is larger than the pixel value of the pixel at which the intersection is located. In this way, it is only necessary to determine that the pixel value of K is smaller than the pixel value of a plurality of pixels in the circumference, and it can be determined that the pixel point where the point K is located is the target pixel point.

S2010. Form, by the determined target pixel, an image of a head region in the image to be identified.

During the recognition process, an image containing a plurality of head regions in the image to be identified may appear, and the images are identified according to the methods of steps S201-S209, and combined according to the determined target pixel points, images of the plurality of head regions may appear. And the statistics of the identified head area image are obtained, and finally the number of people can be obtained.

In the second embodiment of the present invention, the pixel points in the image to be identified are obtained, and according to the arrangement order of the pixel points in the image to be identified, a preset number of pixel points are respectively selected on both sides of the pixel point to form a first pixel point set. And a second pixel point set, the pixel average value of the pixel point in the first pixel point set and the pixel average value of the pixel point in the second pixel point set are subtracted to obtain a difference, if the difference is If the absolute value is less than the preset value, determining that the pixel point is the center point of the preset area, detecting the diameter of the head region in the acquired plurality of image samples, calculating an average value of the diameter, and using the average value as a preset Length, taking the pixel as a center, determining a circular area by using a preset length as a radius, and using the circular area as the preset area, when the pixel value of the pixel is less than or equal to a plurality of pixels in the preset area When the pixel value is determined, the pixel point is determined as a target pixel point, and the determined target pixel point constitutes an image of the head region in the image to be recognized. In this way, it is only necessary to compare the pixel value of a single pixel with the pixel value of a plurality of pixels to determine the pixel of the head region, which is not affected by the clothing, hairstyle and color of the pedestrian, and improves the image recognition of the head region. Precision.

Referring to FIG. 5, FIG. 5 is a schematic structural diagram of an image recognition apparatus according to a third embodiment of the present invention. For convenience of description, only parts related to the embodiment of the present invention are shown. The image recognition device illustrated in FIG. 5 may be an execution body of the image recognition method provided by the foregoing embodiments shown in FIGS. 1 and 2, and may be one of the image recognition device or the image recognition device. The image recognition apparatus illustrated in FIG. 5 mainly includes an acquisition module 31, a determination module 32, and a constituent module 33. The above functional modules are described in detail as follows:

An obtaining module 31, configured to acquire a pixel in the image to be identified;

The determining module 32 is configured to determine that the pixel point is a target pixel point when a pixel value of the pixel point is less than or equal to a pixel value of a plurality of pixel points in the preset area;

The target pixel is a pixel in the head region. The preset area is an area centered on the pixel, and the preset area may be a regular pattern such as a circle, a square, or an ellipse. In an image taken with a depth camera or a somatosensory camera, the pixel value of a pixel of an object is proportional to the distance from the camera to the object. Therefore, the closer the object in the image to be recognized is to the camera, the lower the pixel value of the object. Since the head area is closest to the camera when standing, the pixel value of the head area pixel is the smallest in the image to be recognized, that is, the pixel value of the pixel area of the head area is smaller than the shoulder and the ground waiting Identify pixel values of other parts displayed in the image.

The constituting module 33 is configured to form an image of the head region in the image to be identified by the determined target pixel.

For the details of the present embodiment, please refer to the first embodiment shown in FIG. 1 , and details are not described herein again.

In the third embodiment of the present invention, the obtaining module 31 acquires a pixel in the image to be identified. When the pixel value of the pixel is less than or equal to the pixel value of the plurality of pixels in the preset area, the determining module 32 determines the pixel. And being a target pixel, wherein the preset area is an area centered on the pixel, and the forming module 33 forms an image of the head area in the image to be identified by the determined target pixel. In this way, the pixel value of the single pixel point can be compared with the pixel value of the plurality of pixel points to determine the pixel area of the head region, which is not affected by the clothing, hairstyle and color of the pedestrian, and improves the image recognition of the head region. Precision.

Referring to FIG. 6, FIG. 6 is a schematic structural diagram of an image recognition apparatus according to a fourth embodiment of the present invention. For convenience of description, only parts related to the embodiment of the present invention are shown. The image recognition device illustrated in FIG. 6 may be an execution body of the image recognition method provided by the foregoing embodiment shown in FIGS. 1 and 2, and may be a control module of the image recognition device or the image recognition device. The image recognition apparatus illustrated in FIG. 6 mainly includes an acquisition module 41, a formation module 42, a calculation module 43, a detection module 44, a determination module 45, and a configuration module 46. The above functional modules are described in detail as follows:

An obtaining module 41, configured to acquire a pixel point in the image to be identified;

The forming module 42 is configured to select a preset number of pixel points on both sides of the pixel point according to an arrangement order of the pixel points in the image to be identified, to form a first pixel point set and a second pixel point set;

The calculating module 43 is configured to perform subtraction between the pixel average value of the pixel point in the first pixel point set and the pixel average value of the pixel point in the second pixel point set to obtain a difference value;

a detecting module 44, configured to detect a diameter of a head region in the acquired plurality of image samples;

Further, the calculating module 43 is configured to calculate an average value of the diameter, and use the average value as a preset length;

a determining module 45, configured to determine that the pixel point is a target pixel point when a pixel value of the pixel point is less than or equal to a pixel value of a plurality of pixel points in the preset area;

The preset area is a preset area centered on the pixel. The target pixel is the pixel of the overhead image.

Further, the determining unit 45 is further configured to: if the absolute value of the difference is less than a preset value, determine that the pixel point is a center point of the preset area;

Further, the determining module 45 is further configured to determine a circular area by using the pixel as a center and a preset length as a radius;

Further, the determining module 45 is further configured to use the circular area as the preset area.

The constituting module 46 is configured to form an image of the head region in the image to be identified by the determined target pixel.

For the details of the present embodiment, please refer to the first embodiment shown in FIG. 1 and the second embodiment shown in FIG. 2, and details are not described herein again.

In the fourth embodiment of the present invention, the acquiring module 41 acquires the pixel points in the image to be identified, and the forming module 42 selects a preset number of pixel points on both sides of the pixel according to the order of the pixels in the image to be recognized. Forming a first set of pixel points and a second set of pixel points, and the calculating module 43 performs a subtraction between a pixel average value of the pixel points in the first pixel point set and a pixel average value of the pixel points in the second pixel point set, Obtaining a difference. If the absolute value of the difference is less than the preset value, the determining module 45 determines that the pixel is the center point of the preset area, and the detecting module 44 detects the diameter of the head area in the acquired plurality of image samples. The calculating module 43 calculates an average value of the diameter, and uses the average value as a preset length. The determining module 45 takes the pixel as a center and determines a circular area by using a preset length as a radius. The determining module 45 uses the circular area as the circular area. In the preset area, when the pixel value of the pixel is less than or equal to the pixel value of the plurality of pixels in the preset area, the determining module 45 determines that the pixel is the target pixel, and constitutes Block 46 in the image of the image to be recognized by the head region constituting the target pixel is determined. In this way, it is only necessary to compare the pixel value of a single pixel with the pixel value of a plurality of pixels to determine the pixel of the head region, which is not affected by the clothing, hairstyle and color of the pedestrian, and improves the image recognition of the head region. Precision.

In the various embodiments provided herein, it should be understood that the disclosed systems, devices, and methods may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be another division manner, for example, multiple modules or components may be combined or Can be integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication link shown or discussed may be an indirect coupling or communication link through some interface, device or module, and may be electrical, mechanical or otherwise.

The modules described as separate components may or may not be physically separated. The components displayed as modules may or may not be physical modules, that is, may be located in one place, or may be distributed to multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

The integrated modules, if implemented in the form of software functional modules and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, or all or part of the technical solution, may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: a U disk, a mobile hard disk, a read only memory (ROM, Read-Only) Memory, random access memory (RAM), disk or optical disk, and other media that can store program code.

It should be noted that, for the foregoing method embodiments, for the sake of brevity, they are all described as a series of action combinations, but those skilled in the art should understand that the present invention is not limited by the described action sequence. Because certain steps may be performed in other sequences or concurrently in accordance with the present invention. In the following, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

In the above embodiments, the descriptions of the various embodiments are all focused, and the parts that are not detailed in a certain embodiment can be referred to the related descriptions of other embodiments.

The above is a description of the image recognition method and apparatus provided by the present invention. For those skilled in the art, according to the idea of the embodiment of the present invention, there are changes in the specific implementation manner and application scope. In summary, the present specification The content should not be construed as limiting the invention.

Claims

An image recognition method, the method comprising:

Obtaining pixel points in the image to be identified;

Determining, when the pixel value of the pixel is less than or equal to a pixel value of a plurality of pixel points in the preset area, the pixel point is a target pixel point, wherein the preset area is centered on the pixel point Area;

An image of the head region in the image to be recognized is formed by the determined target pixel.
The method according to claim 1, wherein after the acquiring the pixel points in the image to be identified, the method further comprises:

Selecting a preset number of pixel points on both sides of the pixel to form a first pixel point set and a second pixel point set according to an arrangement order of pixel points in the image to be identified;

Subtracting between a pixel average value of the pixel points in the first pixel point set and a pixel average value of the pixel points in the second pixel point set to obtain a difference value;

If the absolute value of the difference is less than a preset value, it is determined that the pixel point is a center point of the preset area.
The method according to claim 1, wherein when the pixel value of the pixel point is less than or equal to a pixel value of a plurality of pixel points in the preset area, before the pixel point is determined as the target pixel point, Also includes:

Taking the pixel point as a center and determining a circular area by using a preset length as a radius;

The circular area is used as the preset area.
The method according to claim 3, wherein the determining, by using the pixel point as a center and determining the circular area by using the preset length as a radius, further comprises:

Detecting a diameter of a head region in the obtained plurality of image samples;

An average of the diameters is calculated and the average is taken as the preset length.
An image recognition device, characterized in that the device comprises:

Obtaining a module, configured to acquire a pixel in the image to be identified;

a determining module, configured to determine that the pixel point is a target pixel point when a pixel value of the pixel point is less than or equal to a pixel value of a plurality of pixel points in a preset area, where the preset area is The area where the pixel is the center point;

And a component module configured to form an image of the head region in the image to be identified by the determined target pixel.
The device of claim 5, wherein the device further comprises:

a forming module, configured to select a preset number of pixel points on each side of the pixel point according to an arrangement order of pixel points in the image to be identified, to form a first pixel point set and a second pixel point set;

a calculating module, configured to perform a subtraction between a pixel average value of the pixel points in the first pixel point set and a pixel average value of the pixel point in the second pixel point set to obtain a difference value;

The determining module is further configured to: if the absolute value of the difference is less than a preset value, determine that the pixel point is a center point of the preset area.
The device of claim 5 wherein:

The determining module is further configured to determine a circular area by using the pixel as a center and a preset length as a radius;

The determining module is further configured to use the circular area as the preset area.
The device of claim 7 wherein said device further comprises:

a detecting module, configured to detect a diameter of a head region in the obtained plurality of image samples;

The calculation module is further configured to calculate an average value of the diameters, and use the average value as the preset length.