WO2008101422A1

WO2008101422A1 - Image searching method and system

Info

Publication number: WO2008101422A1
Application number: PCT/CN2008/070225
Authority: WO
Inventors: Deng Luo
Original assignee: Tencent Technology (Shenzhen) Company Limited
Priority date: 2007-01-31
Filing date: 2008-01-31
Publication date: 2008-08-28
Also published as: CN101071431A

Abstract

A image searching method and a system are disclosed. The method includes the following steps: obtaining a key graph waiting for matching; analyzing the key graph, obtaining a pattern data of the key graph; determining whether an image in an appointed scope matches with the key graph; obtaining the image matched with the key graph. In this solution, searching the image using the inherent character of the key image as the searching condition, it makes the users searching the needed image quickly and exactly from the plentiful image of the network or the local disk, conveniently the searching operation of the user, saving the searching time of the user, and improving the searching efficiency.

Description

Image search method and system

Technical field

The present invention relates to the field of image search technology, and more particularly to a method and system for image search. Background of the invention

Image technology is a general term for a variety of image-related technologies in a broad sense. According to its characteristics, it can be divided into three levels that are both connected and different: image processing, image analysis and image understanding.

Image processing mainly includes various processing of images to improve the visual effect of the image and to make a cornerstone for automatic recognition; or compression encoding the image, thereby reducing the required storage space or transmission time, and reducing transmission Path requirements.

Image analysis mainly involves detecting and measuring objects of interest in the image to obtain objective information of the target, thereby establishing a description of the image. Image analysis is a process from image to data. The data is used to describe the characteristics and properties of the target in the image, either as a measure of the target feature or as a symbolic representation of the measurement.

Image comprehension is included in the cornerstone of image analysis, further researching the nature of each target in the image and the interrelationship between the targets, and obtaining an understanding of the meaning of the image and an explanation of the original objective scene, thereby guiding And planning image-based actions.

Currently, many search techniques can search for images. For example, google's desktop search engine, microsoft windows disk file search function, etc., can perform image search. Take Windows XP software as an example, manually set the properties of the image file in advance, for example, set the text label of the image file to "My Mountain Travel Photo", or change the name of the image file to "My Mountain Travel Photo", and then search by The software searches for different attributes of the image file and uses "庐山" as a keyword to obtain the target image file. This type of search technology does not search based on the graphical information of the image, but needs to be manually set. Set or machine automatically set, add text or other information to the image, such as text description, camera exif information, aperture, ISO, shutter F value, shooting time, shooting location, author, lens type and other information attached to the map On the image, the information attached to the image is searched.

In the conventional image search method, key information in the form of text or the like is mainly used as a matching condition, and image search is performed in a network or a local computer. That is to say, as the image search matching condition, it is not an image but a text or other type of additional information. These texts or other types of additional information need to be manually attached to the image before they can be used for image search. If an image has no additional text or other type of information, it cannot be searched by the search engine. When the user needs to perform an image search, the additional information of the desired picture is not known. It is usually desirable to use the existing image search method with the image as a matching condition, and it is difficult for the user to search for the desired picture. If the additional information of the image does not fully reflect the characteristics of the image, even if the user searches using the additional information of the image, it is difficult to achieve an accurate and fast image search. Summary of the invention

Embodiments of the present invention provide a method and system for image search that allows a user to quickly and accurately search for a desired image.

A method of image search, comprising:

Get the key graphics to be matched;

Analyzing the key graphic to obtain mode data of the key graphic;

Determining, according to the pattern data of the key graphic, whether an image within the specified range matches the key graphic;

Get an image that matches the key graphic.

A system for image search, comprising: a receiving module, configured to acquire a key graphic to be matched;

An image analysis recording module, configured to analyze the key graphic, and acquire mode data of the key graphic;

And a mode comparison module, configured to determine, according to the pattern data of the key graphic, whether an image within the specified range matches the key graphic, and obtain an image that matches the key graphic.

As can be seen from the above technical solutions, in the embodiment of the present invention, the image is searched using the pattern data of the key pattern as a matching condition. In this way, even if the image has no additional information such as text, the user can quickly and accurately search for the desired image from the network or the local hard disk according to the inherent characteristics of the image, thereby facilitating the user's search operation and saving the user. Search time has greatly improved the efficiency of search. In addition, in the embodiment of the present invention, the key graphics are analyzed by using a graphic analysis technique to obtain pattern data of the key graphics. In this way, image technology is used in the search for images, thereby expanding the field of application of image technology. BRIEF DESCRIPTION OF THE DRAWINGS

1 is a flow chart of a method for image search in an embodiment of the present invention.

2 is a structural diagram of a system for image search in an embodiment of the present invention. Mode for carrying out the invention

In the embodiment of the present invention, the key graphic to be matched is obtained, the key graphic is analyzed, the mode data of the key graphic is obtained, and the image in the specified range is searched according to the mode data of the key graphic, and the key graphic is matched. Image, the image that matches the key graphic is the result of the search.

For a better understanding of the present invention, the method of the present invention will now be described in detail in conjunction with the accompanying drawings.

Figure 1 shows a flow chart of a method of image search in an embodiment. Image search in this embodiment The method of cable includes the following steps.

Step 101: Obtain the key graphic to be matched.

In the embodiment of the present invention, the key graphic has a specific graphic feature, and the specific graphic feature can be used as a matching condition when the image is searched.

In this step, key graphics can be obtained from the drawing area. Users can use the mouse or other input device to draw key graphics on the drawing area that they wish to search.

In the embodiment of the present invention, other methods may also be used to obtain key graphics, such as obtaining from a disk file, scanning through a peripheral device, acquiring through a camera, and the like.

Step 102: Analyze the key graphic to obtain the mode data of the key graphic.

In this step, you can use graphical analysis techniques to analyze key patterns and obtain pattern data for key patterns.

The pattern data of the key graphic may be at least one of the following graphic features: rectangle, circle, triangle, polygon, line, arc, direction, number, color, size, coordinates, human form, and the like.

Step 103: Record the pattern data of the key graphic.

Step 104: Search for the image in the specified range according to the pattern data of the key graphic to determine whether there is an image matching the key graphic.

When searching for an image in this step, comparing the mode data of the image in the specified range with the mode data of the key graphic to determine whether there is an image matching the key graphic, and if so, proceeding to step 105; otherwise, returning none Matching the image, the process of searching for the image ends.

The mode data of the image within the specified range may be preset before the search is performed, or when the search is performed, the mode data of the image is first acquired, and then compared with the mode data of the key graphic.

In this step, if the mode data of the image is at least one of the same as the pattern data of the key graphic, it is determined that the image matches the key graphic, that is, the image is matched with the key graphic. When the key graph includes more than one, you can also set the matching rules as needed.

In the embodiment of the present invention, the matching rule includes: "with matching", "or matching", "no matching". Wherein, if the matching rule is "and match", then only when the image matches each key graphic, the image is determined to match the one or more key graphics; if the matching rule is "or match", as long as the image is at least a key pattern matching, determining that the image matches the one or more key patterns; if the matching rule is "no match", when the image matches at least one of the specified key patterns and does not match the remaining key patterns, It is determined that the image matches the one or more key patterns.

When the embodiment of the present invention is applied, different matching rules may be set according to actual needs, and are not limited to the three matching rules described above.

Step 105: Acquire an image that matches the key graphic.

In this step, an image matching the key pattern is acquired as a result of the image search. Through the above steps, an image matching the key graphic is searched within the specified range based on the pattern data of the key graphic. Since the pattern data of the key pattern is used as the matching condition, that is, the search is performed according to the characteristics inherent to the image, the image can be searched without setting additional information for the image, thereby facilitating the user's search operation and saving The user's search time has improved the efficiency of the search.

Embodiments of the present invention also provide a system for image search. Fig. 2 is a view showing the configuration of an image search system in the embodiment of the present invention. The system includes: a receiving module 201, an image analysis recording module 202, and a mode comparison module 203.

The receiving module 201 is configured to acquire key graphics to be matched.

The receiving module 201 can include sub-modules for obtaining key graphics from the drawing area. The user can use the mouse or other input device to draw the key graphics that he or she wants to search on the drawing area. The receiving module 201 may also include a submodule for supporting other ways of acquiring key graphics. Blocks, for example, support sub-modules for obtaining key graphics from disk files, sub-modules that support scanning through peripherals, sub-modules that support key graphics through the camera, and so on.

The image analysis recording module 202 is configured to analyze the key graphic, acquire the mode data of the key graphic, and record the mode data.

The image analysis and recording module 202 specifically includes an image analysis module 2021 and a graphic feature recording module 2022.

The image analysis module 2021 is configured to analyze the pattern data of the key graphic to obtain the mode data of the key graphic.

The graphic feature recording module 2022 is configured to record the mode data analyzed by the image analysis module. The image analysis recording module 202 may further include sub-modules for setting pattern data of an image within a specified range before the search is performed, or sub-modules for acquiring pattern data of an image within a specified range while the search is in progress.

The mode comparison module 203 is configured to determine, according to the pattern data of the key graphic, whether the image within the specified range matches the key graphic, and obtain an image that matches the key graphic.

The method for graphic search in the embodiment of the present invention is further described below in conjunction with specific embodiments.

The acquired mode data of XI and X2 are shown in Table 1. Image XI image X2

Rectangular 4 0

Rectangular side length (percent) 40, 50 0

Rectangle coordinates (percent) 30,70 0

Rectangular color RGB(34,60,30) 0

Round 0 1

Radius of the circle (percent) 0 80 Circular coordinates (percent) 0 10,10 Circular color 0 RGB(200,10,50) Triangle 0 0

Triangle vertex coordinates (percent) 0 0

Triangle color 0 0

Polygon 0 0

Polygon vertex coordinates (percent) 0 0

Polygon color 0 0

Number of lines 3 0

Straight line direction (angle) 38,57,149 0

Arc 1 0

Arc Radius (%) 25 0

Human form 0 1

Mode data for XI and X2.

The user draws a rectangle in the drawing area, that is, the graphic Y, as the key graphic. After the receiving module acquires the rectangle, the image analyzing module analyzes the key graphic to obtain the mode data of the key graphic, as shown in Table 2. User input graphic Y

Rectangle 1

Rectangular side length (percent) 80, 60

Rectangle coordinates (percent) 20,50

Rectangular color RGB(56,160,130) Circle 0

Radius of the circle (percent) 0

Circular coordinates (percent) 0 Round color 0

Triangle 0

Triangle vertex coordinates (percent) 0

Triangle color 0

Polygon 0

Polygon vertex coordinates (percent) 0

Polygon color 0

Number of lines 0

Straight line direction (angle) 0

Arc 0

Arc Radius (%) 0

Humanoid 0

Table 2 The image feature recording module stores the pattern data of the key figures, and combines the pattern data shown in Table 1 and Table 2, as shown in Table 3. Image XI image X2 user input graphic Y rectangle 4 0 1

Rectangular side length (percent) 40, 50 0 80, 60

Rectangular coordinates (percent) 30,70 0 20,50

Rectangular color RGB(34,60,30) 0 RGB(56,160,130) Circle 0 1 0

Radius of the circle (percent) 0 80 0

Circular coordinates (percent) 0 10,10 0

Round color 0 RGB(200,10,50) 0

Triangle 0 0 0

Triangle apex sits

0 0 0

Standard (percentage) Triangle color 0 0 0 polygon 0 0 0

Polygon vertices sit

0 0 0

Standard (percentage)

Polygon color 0 0 0

Number of lines 3 0 0

Straight line direction (angle) 38,57,149 0 0

Arc 1 0 0

Arc Radius (%) 25 0 0

Human form 0 1 0

The mode comparison module compares the pattern data of the key pattern Y with the pattern data of the images XI and X2 within the specified range to determine whether XI and X2 match the key pattern Y;

It can be seen from Table 3 that the pattern data of the key graphs Y and XI has the same one, that is, both rectangles are included, and the graphs Y and XI match. Although the number of rectangles, coordinates, and colors are not the same, if at least one of the pattern data of XI and Y is the same, it is determined that XI and Y match. If the user draws two key figures Y1 and Y2 in the drawing area, similarly, the pattern data for each key figure is shown in Table 4. User input chart user input graphic image XI image X2

Shape Y1 Y2

Rectangular 4 0 1 0

Rectangular side length

40, 50 0 80, 60 0 (percentage)

Rectangular coordinates

30,70 0 20,50 0 (percentage)

Rectangular color RGB (34,60, 0 RGB (56,160, 0 30) 130)

Round 0 1 0 1

Circular radius

0 80 0 68 (percentage)

Circular coordinates

0 10,10 0 23,62

(percentage)

RGB (200, 10, RGB (255, 255, round color 0 0)

50) 255)

Triangle 0 0 0 0

Triangle

Vertex coordinates 0 0 0 0

(percentage)

Triangle

0 0 0 0 color

Polygon 0 0 0 0

Polygon

Vertex coordinates 0 0 0 0

(percentage)

Polygon

0 0 0 0 color

Number of straight lines 3 0 0 0

Straight line direction

38,57,149 0 0 0

(angle)

Arc 1 0 0 0

Arc radius

25 0 0 0 (percentage)

Human form 0 1 0 1

Table 4 In this embodiment, the user can specify a matching rule as needed.

When the user-specified matching rule is "and match", that is, the image matches Y1 and Y2 at the same time, it is determined that the image matches the key pattern input by the user. It can be seen from Table 4 that XI matches Y1, and X2 matches Y2, that is, XI and X2 do not match the key figures Y1 and Y2 at the same time, that is, according to "match", no key pattern matching with user input is searched. When the user-specified matching rule is "or match", that is, when the image matches at least one of Y1 and Y2, it is determined that the image matches the key pattern input by the user. As can be seen from Table 4, XI matches Y1, and X2 matches Y2, that is, according to "or match", the pattern matching the key pattern input by the user includes XI and X2.

When the user-specified matching rule is "No Match", that is, when the image matches the specified key graphic and does not match the remaining key graphic, it is determined that the image matches the key graphic input by the user. From Table 4, XI only matches Y1, X2 only matches Y2, that is, if in the "No Match", the user-specified key pattern to be matched is Y1, then XI is matched with the key pattern input by the user. Image; if in the "No Match", the user-specified key pattern to be matched is Y2, then X2 is the image that matches the key pattern entered by the user.

If the user enters multiple key figures, you can also set matching rules as needed to determine which images match the key figures.

It can be seen from the above detailed description of the method and system for image search provided by the embodiment of the present invention that since the pattern data of the key graphic is used as the matching condition in the embodiment of the present invention, the image does not need to be provided with additional text or Other types of information can search the image only according to the characteristics inherent in the image, which facilitates the user's search operation, saves the user's search time, and improves the search efficiency.

The description of the above embodiments is only for helping to understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of the present invention, in the specific implementation The scope of the present application and the scope of the application are subject to change. In the above description, the contents of the present specification should not be construed as limiting the invention.

Claims

Claim

A method for image search, comprising:

Get the key graphics to be matched;

Analyzing the key graphic to obtain mode data of the key graphic;

Get an image that matches the key graphic.

2. The method of claim 1, wherein the obtaining the key graphic comprises: acquiring a graphic drawn by the user in a preset drawing area.

3. The method of claim 1, wherein the pattern data comprises at least one of a rectangle, a circle, a triangle, a polygon, a line, an arc, a direction, a quantity, a color, a size, a coordinate, and a figure.

The method according to claim 1, wherein the determining whether the image within the specified range matches the key graphic comprises:

Based on the pattern data of the key pattern, it is determined whether the pattern data of the key pattern and the pattern data of the image set in advance match.

Obtaining pattern data of the image;

Based on the pattern data of the key pattern, it is determined whether the pattern data of the key pattern and the pattern data of the image match.

The method according to claim 4 or 5, wherein the determining the pattern data of the key graphic and the pattern data matching of the image comprises:

When the pattern data of the image is at least one of the pattern data of the key graphic, It is determined that the image matches the key pattern.

The method according to claim 6, wherein the key pattern is more than one; the determining that the image within the specified range matches the key pattern comprises:

Determining that the image matches the one or more key patterns when each of the key patterns matches the image; or

Determining that the image matches the one or more key patterns when at least one of the key graphics matches the image; or

When the specified at least one key graphic matches the image, and the other key graphics do not match the image, it is determined that the image matches the one or more key graphics.

8. A system for image search, comprising:

a receiving module, configured to acquire a key graphic to be matched;

9. The system of claim 8, wherein the receiving module comprises: a sub-module for acquiring the key graphic from a drawing area.

10. The system of claim 8, wherein the image analysis recording module comprises:

An image analysis module, configured to analyze the key graphic, and acquire mode data of the key graphic;

A graphic feature recording module is configured to store mode data of the key graphic.

11. The system of claim 8, wherein the image analysis recording module further comprises:

A sub-module for acquiring pattern data of the image.