CN116127118A - Method, device, electronic equipment and storage medium for searching similar images - Google Patents

Abstract

The application discloses a method, a device, electronic equipment and a storage medium for similar image retrieval, and belongs to the technical field of image processing. The method comprises the following steps: determining a first feature of a target image and a second feature of the target image, wherein the first feature is used for describing the integral feature of the image, and the second feature is used for describing the local feature of the image; determining at least one image to be selected from an image database, wherein the first characteristic of the image to be selected is the same as the first characteristic of the target image; and determining a second image similar to the target image from the at least one image to be selected according to the second characteristic of the target image and the second characteristic of the at least one image to be selected.

Description

A method, device, electronic equipment and storage medium for similar image retrieval

technical field

The present application belongs to the technical field of image processing, and in particular relates to a similar image retrieval method, device, electronic equipment and storage medium.

Background technique

In the field of image processing technology, it is often necessary to retrieve similar images from large-scale images for analysis. The related method is to first generate the fingerprint f0 of the target image, which is generally a fixed-length string or a fixed-length floating-point number vector, and then calculate the distance between f0 and all fingerprints in the image fingerprint library, and the image corresponding to the fingerprint with the smallest distance is It is an image similar to the target image, but due to the huge image fingerprint database, each retrieval takes a long time and the efficiency is low.

Contents of the invention

The purpose of the embodiments of the present application is to provide a similar image retrieval method, device, electronic device, and storage medium, which can improve the efficiency of similar image retrieval.

In order to solve the above-mentioned technical problems, the application is implemented as follows:

In the first aspect, the embodiment of the present application provides a method for similar image retrieval, the method includes: determining the first feature of the target image and the second feature of the target image, the first feature is used to describe the overall image feature, the second feature is used to describe the local features of the image; from the image database, determine at least one candidate image, wherein the first feature of the candidate image is the same as the first feature of the target image; according to The second feature of the target image and the second feature of the at least one candidate image determine a second image similar to the target image from the at least one candidate image.

In the second aspect, the embodiment of the present application provides a device for similar image retrieval, the device includes: a first determination module, configured to determine the first feature of the target image and the second feature of the target image, the first The feature is used to describe the overall feature of the image, and the second feature is used to describe the local feature of the image; the screening module is used to determine at least one candidate image from the image database, wherein the first feature of the candidate image The same as the first feature of the target image; the second determination module is configured to select from the at least one candidate image according to the second feature of the target image and the second feature of the at least one candidate image, A second image similar to the target image is determined.

In the third aspect, the embodiment of the present application provides an electronic device, the electronic device includes a processor and a memory, the memory stores programs or instructions that can run on the processor, and the programs or instructions are processed by the The steps of the method for retrieving similar images as described in the first aspect are realized when the device is executed.

In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, similar image retrieval as described in the first aspect is realized steps of the method.

In the fifth aspect, the embodiment of the present application provides a chip, the chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run programs or instructions, so as to implement the first aspect The method of similar image retrieval described above.

In a sixth aspect, an embodiment of the present application provides a computer program product, the program product is stored in a storage medium, and the program product is executed by at least one processor to implement the similar image retrieval method as described in the first aspect.

In the embodiment of the present application, by determining the first feature of the target image and the second feature of the target image, the first feature is used to describe the overall feature of the image, and the second feature is used to describe the local feature of the image ; From the image database, determine at least one candidate image, wherein the first feature of the candidate image is the same as the first feature of the target image; according to the second feature of the target image and the at least one candidate image The second feature of the selected image is used to determine a second image similar to the target image from the at least one candidate image, which can improve the efficiency of similar image retrieval.

Description of drawings

FIG. 1 shows a schematic flow chart of a method for similar image retrieval provided in an embodiment of the present application;

Fig. 2 shows another schematic flow chart of the similar image retrieval method provided by the embodiment of the present application;

FIG. 3 shows a schematic structural diagram of a device for similar image retrieval provided by an embodiment of the present application;

FIG. 4 shows a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present application.

Detailed ways

The following will clearly describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, but not all of them. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in this application belong to the protection scope of this application.

The terms "first", "second" and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific sequence or sequence. It should be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the application can be practiced in sequences other than those illustrated or described herein, and that references to "first," "second," etc. distinguish Objects are generally of one type, and the number of objects is not limited. For example, there may be one or more first objects. In addition, "and/or" in the specification and claims means at least one of the connected objects, and the character "/" generally means that the related objects are an "or" relationship.

The method for retrieving similar images provided by the embodiment of the present application will be described in detail below through specific embodiments and application scenarios with reference to the accompanying drawings.

Fig. 1 is a schematic flowchart of a method for similar image retrieval provided by an embodiment of the present application. As shown in Fig. 1 , the method may include the following steps.

S101: Determine a first feature of a target image and a second feature of the target image.

Wherein, the first feature is used to describe the overall feature of the image, and the second feature is used to describe the local feature of the image. Specifically, the first feature includes, but is not limited to, a color feature and a texture feature of an image. The second feature may be a contour feature in the image, a spatial relationship feature, or a feature of a part of the image, that is, the image is divided into multiple partial areas, and the second feature may be a feature of at least one partial area.

Optionally, the feature may be expressed in at least one manner such as character sequence, vector, matrix, or function expression.

S102: Determine at least one candidate image from the image database.

Wherein, the first feature of the candidate image is the same as the first feature of the target image. That is, the image to be selected has the same overall characteristics as the target image. The image database includes multiple images of the same or similar category or the same source as the target image. For example, the target image is a remote sensing image, and the image database contains multiple images belonging to the category of remote sensing images or natural landscape images.

In this step, there is no need to calculate the relevant similarity distance, and a small number of candidate images that are the same as the first feature of the target image can be determined from a large image database, which reduces the scope of retrieval and avoids the related similarity. The problem of large consumption and low efficiency of distance computing resources saves computing resources and improves computing efficiency.

S103: Determine a second image similar to the target image from the at least one candidate image according to the second feature of the target image and the second feature of the at least one candidate image.

Specifically, the area described by the second feature of the target image corresponds to the area described by the second feature of the candidate image at least in size and position. For example, a second feature of the target image is used to describe the upper left part of the target image with a width of 100px (pixel, pixel) and a height of 130px. Correspondingly, the corresponding second feature of the image to be selected is also used to describe the selected The upper left part of the image to be selected has a width of 100px and a height of 130px.

In this step, among the determined small number of candidate images, a second image similar to the target image is further determined according to the second feature of the target image and the second feature of the at least one candidate image. Because the number of images to be selected is much smaller than the number of images in the image database, the efficiency of calculation and retrieval is greatly improved.

In this embodiment of the present application, by determining the first feature of the target image and the second feature of the target image, the first feature is used to describe the overall feature of the image, and the second feature is used to describe the local feature of the image; from the image In the database, at least one candidate image is determined, wherein the first feature of the candidate image is the same as the first feature of the target image; according to the second feature of the target image and the at least one candidate image The second feature is to determine a second image similar to the target image from the at least one candidate image, which can improve the efficiency of similar image retrieval.

Fig. 2 is another schematic flowchart of a similar image retrieval method provided in the embodiment of the present application. As shown in Fig. 2 , the method may include the following steps.

S201: Determine a feature of each channel of the target image according to multiple pixel values of each channel of the target image; determine a first feature of the target image according to the features of each channel.

Specifically, the channels include but are not limited to three channels of red, green and blue RGB, and the value range of the pixel value of each channel is an integer from 0 to 255. Or three CMY channels composed of cyan, magenta and yellow, and the value range of the pixel value of each channel is an integer from 0 to 100. The feature of each channel may be represented by at least one of vector, matrix or expression.

Optionally, the range of pixel values of the channel is reduced, and the representation of the feature of the channel is simplified without changing the feature of the target image. For example, a pixel value r _ij of the R channel is converted by the following formula:

表示如果r_ij是4的整数倍，则不改变r_ij的像素值；如果r_ij不是4的整数倍，则减小r_ij，直至其为4的整数倍。通过该公式可以将邻近的像素值用一个像素值表示，即将像素值的取值范围从0～255中的整数缩小至0～252中4的整数倍。Among them, mod4 means that r _ij performs a remainder operation on 4,

It means that if r _ij is an integer multiple of 4, the pixel value of r _ij will not be changed; if r _ij is not an integer multiple of 4, then reduce r _ij until it is an integer multiple of 4. Through this formula, adjacent pixel values can be represented by one pixel value, that is, the value range of the pixel value is reduced from an integer in 0 to 255 to an integer multiple of 4 in 0 to 252.

Through the above-mentioned processing of image pixel values, that is, converting two pixels whose pixel values differ within 4 into the same pixel (classified into one category), the generated first feature is more robust.

S202: Determine at least one first image from the target image.

Wherein, the first image is an image of a part of the target image, and different first images correspond to different parts of the target image. Specifically, the part in the target image corresponding to the at least one first image may be a regular part or an irregular part. For example, the rectangular target image is equally divided into four small rectangular first images according to equal height and width, or other division methods of unequal division, which can be set according to actual conditions, and are not specifically limited here.

Optionally, there may also be overlapping parts between the first images.

S203: For each of the first images, according to a plurality of pixel values of each channel of the first image, determine the characteristics of each channel of the first image; according to each channel of the first image and determining a second feature of the target image corresponding to the first image.

This step can adopt the description of step S201 in this embodiment, and will not be repeated here.

It should be noted that, the more the above-mentioned determined first images and the larger the range of the target image covered by the plurality of first images, the more second features are extracted and the feature expression is richer.

S204: Determine at least one candidate image from the image database.

This step can adopt the description of step S102 in the embodiment of FIG. 1 , and will not be repeated here.

S205: For each candidate image, respectively compare each second feature of the target image with the corresponding second feature of the candidate image, and determine the candidate image according to multiple comparison results. The similarity value corresponding to the selected image.

Specifically, each second feature of the target image is compared with the corresponding second feature of the candidate image for similarity, and the result of the comparison may be the result of scoring according to the similarity between each corresponding second feature, Score values can be expressed as decimals or percentages between 0 and 1. For example, if the target image and the candidate image have 4 sets of corresponding second features, that is, there are 4 corresponding first images, the following formula can be used for scoring:

表示待选图像p_i在第j个第二特征上的得分值，

表示待选图像p_i的第j个第二特征，

表示目标图像p₀的第j个第二特征。where j=1,2,3,4,

Indicates the score value of the candidate image p _i on the jth second feature,

Indicates the jth second feature of the candidate image p _i ,

Denotes the jth second feature of the target image _p0 .

Then, the similarity value between the candidate image and the target image is determined according to the comparison results of multiple groups corresponding to the second feature. Specifically, the comparison results of the plurality of groups corresponding to the second feature may be summed or averaged to obtain the similarity value between the candidate image and the target image. For example, the following formula is used to sum the results of each group of second feature comparisons as the similarity value of p _i and p ₀ :

Optionally, in order to differentiate the similarity results between different candidate images, the following formula is used to increase the comparison of a set of second features (the fifth) to achieve the accuracy of the comparison results:

和

表示目标图像p₀的第二特征在三个颜色通道的特征分量，对in,

and

Represents the feature components of the second feature of the target image p ₀ in the three color channels, for

和

表示待选图像p_i的第二特征在三个颜色通道的特征分量。对三个颜色通道的特征分量一一对应比较，选出差异最大的特征分量，进而计算该特征分量对应颜色通道上p₀和p_i的相似度。should,

and

Represents the feature components of the second feature of the candidate image p _i in the three color channels. The feature components of the three color channels are compared one by one, and the feature component with the largest difference is selected, and then the similarity between p ₀ and p _i on the corresponding color channel of the feature component is calculated.

It should be noted that, the above-mentioned similarity score value corresponding to each second feature is set to 0.2, so that the similarity value of the finally obtained candidate image is between [0,1], but the score value can also be Set to other values, when the score value is other values, the final similarity value can be normalized to [0,1].

S206: Determine the candidate image with the highest similarity value as the second image.

Specifically, before this step, the similarity values of each candidate image are sorted to determine the candidate image with the highest similarity value. In an implementation manner, the determining the first feature of the target image includes: determining the following character string f ₀ as the first feature of the target image:

f ₀ =h×w,δ ₁ _δ ₂ ,j ₁₁ _j ₁₂ _…_j _1m _j ₂₁ _j ₂₂ _…_j _2m _j ₃₁ _j ₃₂ _…_j _3m , where h and w are the heights of the target image and width, δ ₁ , δ ₂ and m are preset parameters, satisfying δ ₁ >δ ₂ , m∈[2,255], j _i1 , j _i2 , ... j _im are the multi- Among the pixel values, the grayscale ratio is greater than or equal to the first m pixel values of δ ₁ , and δ ₂ is used to adjust the order of j _i1 , j _i2 , ... j _im , and the i-th channel includes red, green, blue, RGB one of the channels.

将每个通道中像素值为j的灰度所占的比例u_j从大到小排序，截取灰度占比大于等于δ₁的前m个比值组成向量，即

即满足条件

k＝1,2,...,n，如果大于等于δ₁的比值u_j不足m个，那么用u_-1＝0补足。例如，u＝

下标(j_i1，j_i2，-1,-1,-1)对应的像素值为(252，68，-1，-1，-1)，对于每个通道的像素都做类似的处理，即可得到所述第一特征。Specifically, the grayscale proportions corresponding to the multiple pixel values contained in each channel are sorted from large to small, and the first m grayscale proportions greater than or equal to _δ1 are intercepted to form a feature vector, for example, _δ1 =1. After calculating the gray scale ratio of each pixel value, the

Sort the proportion u _j of the grayscale of pixel value j in each channel from large to small, and intercept the first m ratios whose grayscale proportion is greater than or equal to δ ₁ to form a vector, that is

That is to meet the conditions

k=1,2,...,n, if there are less than m ratios u _j greater than or equal to δ ₁ , use u ₋₁ =0 to make up. For example, u =

The pixel value corresponding to the subscript (j _i1 , j _i2 , -1,-1,-1) is (252, 68, -1, -1, -1), and similar processing is done for the pixels of each channel. The first feature can be obtained.

In order to make the extracted features have better generalization ability, this embodiment can also adjust the above-mentioned extracted feature vectors. Specifically, if the proportion of adjacent gray levels satisfies the result of the difference between the two is less than δ ₂ , and When the grayscale ratio of the corresponding high pixel value is before the grayscale ratio of the corresponding low pixel value, where δ ₂ can be set to a smaller number smaller than _δ1 , then the grayscale ratio of the low pixel value The order of the grayscale ratio with the higher pixel value in the feature vector is exchanged. For example, the eigenvector

u=(7.058, 6.922, 5.824, 4.913, 4.518

The corresponding pixel values are (12, 16, 20, 24, 8), take δ ₂ =0.5 for sequential adjustment, and get the adjusted feature vector

The corresponding pixel values are (12, 16, 20, 8, 24).

For example, the feature height and width of the target image are h=421 and w=690 respectively, and the preset parameters δ ₁ =1.0, δ ₂ =0.5, m=5, then the first feature of the target image can be expressed as:

f ₀ =421×690,1.0_0.5,0_4_8_12_16_0_4_8_12_16_0_4_8_12_16.

和

的所有元素都是0，那么令δ₁＝δ₁/2，然后重新计算

和

指纹f₀是一个用逗号分割的字符串，共包含3段。当m＝5时，f₀＝h×w,δ₁_δ₂,j₁₁_j₁₂_j₁₃_j₁₄_j₁₅_j₂₁_j₂₂_j₂₃_j₂₄_j₂₅_j₃₁_j₂₂_j₃₃_j₃₄_j₃₅，其中×是字符，_也是字符。Additionally, if

and

All elements of are 0, then let δ ₁ = δ ₁ /2, and then recalculate

and

Fingerprint f ₀ is a comma-separated character string consisting of 3 segments. When m=5, f ₀ =h×w,δ ₁ _δ ₂ ,j ₁₁ _j ₁₂ _j ₁₃ _j ₁₄ _j ₁₅ _j ₂₁ _j ₂₂ _j ₂₃ _j ₂₄ _j ₂₅ _j ₃₁ _j ₂₂ _j ₃₃ _j ₃₄ _j ₃₅ , where × is a character, and _ is also a character.

In an implementation manner, the second feature of the target image includes: determining the following character string f _k as the second feature of the target image:

f _k =l ₁₁ _l ₁₂ _…_l _1m _l ₂₁ _l ₂₂ _…_l _2m _l ₃₁ _l ₃₂ _…_l _3m

或h₁＝h-

所述第一图像的宽

或

或

或

h和w分别为所述目标图像的高和宽。Among them, l _i1 , l _i2 , ..., l _im are the first m pixel values whose grayscale ratio is greater than or equal to δ ₁ among the plurality of pixel values of the i-th channel of the first image, and the i-th channel Including one of the three channels of red, green and blue RGB, the first image is the image of the kth part of the target image, k=1, 2,..., K, K is the first image in the target image The number of the first image of the high

or h ₁ =h-

The width of the first image

or

or

or

h and w are the height and width of the target image, respectively.

Wherein, h and w are the height and width of the target image respectively. The above settings for the first image can effectively retrieve similar images whose only difference lies in whether there is a watermark. Moreover, the features of the image are represented by character strings, and by comparing whether two character strings are the same, no distance calculation is required, which saves computing resources and improves computing efficiency.

Optionally, the gray ratio of each pixel value can also be used as the second feature of the target image, so as to further enrich the extracted features and improve the accuracy of image retrieval. In an implementation manner, the second feature of the target image further includes: determining the following vectors as the second feature of the target image:

分别为所述目标图像的第i个通道中像素值分别为j_i1、j_i2、…、j_im的灰度占比，

为所述目标图像的第i个通道中多个像素值中像素值等于j_im的个数。in,

are respectively the grayscale ratios of pixel values j _i1 , j _i2 , ..., j _im in the ith channel of the target image,

is the number of pixel values equal to j _im among multiple pixel values in the i-th channel of the target image.

In this embodiment of the present application, the feature of each channel of the target image is determined according to multiple pixel values of each channel of the target image; and the first feature of the target image is determined according to the feature of each channel ; Determining at least one first image from the target image; For each of the first images, according to a plurality of pixel values of each channel of the first image, determine the value of each channel of the first image feature; according to the feature of each channel of the first image, determine the second feature of the target image corresponding to the first image; from the image database, determine at least one candidate image; for each selected the image to be selected, respectively comparing each second feature of the target image with the corresponding second feature of the image to be selected, and determining the similarity corresponding to the image to be selected according to a plurality of comparison results value; determining the candidate image with the highest similarity value as the second image can save computing resources and improve the efficiency and accuracy of similar image retrieval.

FIG. 3 shows a schematic structural diagram of an apparatus for similar image retrieval provided by an embodiment of the present application. The apparatus 300 includes: a first determination module 301 , a screening module 302 and a second determination module 303 .

The first determining module 301 is used to determine a first feature of the target image and a second feature of the target image, the first feature is used to describe the overall feature of the image, and the second feature is used to describe the local feature of the image; The screening module 302 is configured to determine at least one candidate image from the image database, wherein the first feature of the candidate image is the same as the first feature of the target image; the second determination module 303 is configured to The second feature of the image and the second feature of the at least one candidate image determine a second image similar to the target image from the at least one candidate image.

In an implementation manner, the second determining module 303 is further configured to, for each of the candidate images, respectively compare each second feature of the target image with the corresponding second feature of the candidate image Comparing, according to multiple comparison results, determining the similarity value corresponding to the candidate image; determining the candidate image with the highest similarity value as the second image.

In an implementation manner, the first determining module 301 is further configured to determine the feature of each channel of the target image according to multiple pixel values of each channel of the target image; The features of the first feature of the target image are determined.

In an implementation manner, the first determining module 301 is further configured to determine at least one first image from the target image, the first image is a part of the target image, different first images The parts of the corresponding target images are different; for each of the first images, according to a plurality of pixel values of each channel of the first image, the characteristics of each channel of the first image are determined; according to the The feature of each channel of the first image determines the second feature of the target image corresponding to the first image.

In an implementation manner, the first determination module 301 is further configured to determine the following character string f ₀ as the first feature of the target image:

f ₀ =h×w,δ ₁ _δ ₂ ,j ₁₁ _j ₁₂ _…_j _1m _j ₂₁ _j ₂₂ _…_j _2m _j ₃₁ _j ₃₂ _…_j _3m

Wherein, h and w are the height and width of the target image respectively, δ ₁ , δ ₂ and m are preset parameters, satisfying δ ₁ >δ ₂ , m∈[2,255], j _i1 , j _i2 ,..., j _im is the first m pixel values whose grayscale ratio is greater than or equal to _δ1 among the plurality of _pixel values of the i-th channel of the target image, and the i-th channel includes one of the three channels of red, green, blue, RGB By.

In an implementation manner, the first determining module 301 is further configured to determine the following character string f _k as the second feature of the target image:

f _k = ₁₁ _ ₁₂ _…_l _1m _ ₂₁ _ ₂₂ _…_l _2m _ ₃₁ _ ₃₂ _…_l _3m

或

所述第一图像的宽

或

或

或

h和w分别为所述目标图像的高和宽。Among them, l _i1 , l _i2 , ..., l _im are the first m pixel values whose grayscale ratio is greater than or equal to δ ₁ among the plurality of pixel values of the i-th channel of the first image, and the i-th channel Including one of the three channels of red, green and blue RGB, the first image is the image of the kth part of the target image, k=1, 2,..., K, K is the first image in the target image The number of the first image of the high

or

The width of the first image

or

or

or

h and w are the height and width of the target image, respectively.

In an existing manner, the first determining module 301 is further configured to determine the following vectors as the second feature of the target image:

分别为所述目标图像的第i个通道中像素值分别为j_i1、j_i2、…、j_im的灰度占比，

为所述目标图像的第i个通道中多个像素值中像素值等于j_im的个数。in,

are respectively the grayscale ratios of pixel values j _i1 , j _i2 , ..., j _im in the ith channel of the target image,

is the number of pixel values equal to j _im among multiple pixel values in the i-th channel of the target image.

The device 300 provided in the embodiment of the present application can execute the methods described in the foregoing method embodiments, and realize the functions and beneficial effects of the methods described in the foregoing method embodiments, which will not be repeated here.

Fig. 4 shows a schematic diagram of the hardware structure of an electronic device performing similar image retrieval provided by an embodiment of the present application. Referring to this figure, at the hardware level, the electronic device includes a processor, and optionally includes an internal bus, a network interface, and a memory. Wherein, the memory may include a memory, such as a high-speed random-access memory (Random-Access Memory, RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory. Of course, the electronic device may also include hardware required by other services.

The processor, network interface, and memory can be interconnected by an internal bus, which can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an extended industry standard Structure (Extended Industry Standard Architecture, EISA) bus, etc. The bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one double-headed arrow is used in this figure, but it does not mean that there is only one bus or one type of bus.

Memory for storing programs. Specifically, the program may include program code, and the program code includes computer operation instructions. Storage, which can include internal memory and nonvolatile storage, provides instructions and data to the processor.

The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a device for locating target users on a logical level. The processor executes the program stored in the memory, and is specifically used to execute the methods described in Figure 1-2 in the foregoing method embodiments, and realize the functions and benefits of the various methods described in Figure 1-2 in the foregoing method embodiments effect, which will not be repeated here.

The above method disclosed in the embodiments shown in FIGS. 1-2 of the present application may be applied to a processor or implemented by the processor. A processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software. The above-mentioned processor can be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processor, DSP), a dedicated integrated Circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. Various methods, steps, and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register. The storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps of the above method in combination with its hardware.

The electronic device may also execute the methods described in FIGS. 1-2 in the foregoing method embodiments, and realize the functions and beneficial effects of the methods described in FIGS. 1-2 in the foregoing method embodiments, which will not be repeated here.

Of course, in addition to the software implementation, the electronic device of the present application does not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is to say, the execution subject of the following processing flow is not limited to each logic unit, It can also be a hardware or logic device.

The embodiment of the present application also provides a computer-readable storage medium, the computer-readable medium stores one or more programs, and when the one or more programs are executed by an electronic device including multiple application programs, the The electronic device executes the methods described in FIGS. 1-2 in the foregoing method embodiments, and realizes the functions and beneficial effects of the methods described in FIGS. 1-2 in the foregoing method embodiments, and details are not repeated here.

Wherein, the computer-readable storage medium includes a read-only memory (Read-Only Memory, ROM for short), a random access memory (Random Access Memory, RAM for short), a magnetic disk or an optical disk, and the like.

Further, the embodiment of the present application also provides a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions When executed by a computer, the functions and beneficial effects of the methods described in Figs. 1-2 in the foregoing method embodiments are realized, and details are not repeated here.

In a word, the above descriptions are only preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this application shall be included within the protection scope of this application.

The systems, devices, modules, or units described in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementing device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or Combinations of any of these devices.

Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can be implemented by any method or technology for storage of information. Information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Flash memory or other memory technology, Compact Disc Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cartridge, tape magnetic disk storage or other magnetic storage device or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media excludes transitory computer-readable media, such as modulated data signals and carrier waves.

It should be noted that, in this document, the term "comprising", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising that element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved. Functions are performed, for example, the described methods may be performed in an order different from that described, and various steps may also be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is better implementation. Based on such an understanding, the technical solution of the present application can be embodied in the form of computer software products, which are stored in a storage medium (such as ROM/RAM, magnetic disk, etc.) , optical disc), including several instructions to enable a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present application.

The embodiments of the present application have been described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementations. The above-mentioned specific implementations are only illustrative and not restrictive. Those of ordinary skill in the art will Under the inspiration of this application, without departing from the purpose of this application and the scope of protection of the claims, many forms can also be made, all of which belong to the protection of this application.

Claims (10)

1. A method of similar image retrieval, the method comprising:

determining a first feature of a target image and a second feature of the target image, wherein the first feature is used for describing the integral feature of the image, and the second feature is used for describing the local feature of the image;

determining at least one image to be selected from an image database, wherein the first characteristic of the image to be selected is the same as the first characteristic of the target image;

and determining a second image similar to the target image from the at least one image to be selected according to the second characteristic of the target image and the second characteristic of the at least one image to be selected.

2. The method of claim 1, wherein the second features comprise a plurality of features for describing different portions of the image;

the determining, from the at least one candidate image, a second image similar to the target image according to the second feature of the target image and the second feature of the at least one candidate image, includes:

comparing each second characteristic of the target image with the corresponding second characteristic of the image to be selected according to each image to be selected, and determining a similarity value corresponding to the image to be selected according to a plurality of comparison results;

and determining the image to be selected with the highest similarity value as the second image.

3. The method of claim 1, wherein determining the first characteristic of the target image comprises:

determining a characteristic of each channel of the target image according to a plurality of pixel values of each channel of the target image;

and determining a first characteristic of the target image according to the characteristic of each channel.

4. The method of claim 1, wherein determining the second characteristic of the target image comprises:

determining at least one first image from the target images, wherein the first image is an image of part of the target images, and the parts of the target images corresponding to different first images are different;

determining, for each of the first images, a feature of each channel of the first image from a plurality of pixel values of each channel of the first image; and determining a second characteristic of the target image corresponding to the first image according to the characteristic of each channel of the first image.

5. A method according to claim 1 or 3, wherein said determining a first characteristic of the target image comprises:

the following character string f will be described ₀ Determining a first feature of the target image:

f ₀ ＝h×w，δ ₁ _δ ₂ ,j ₁₁ _j ₁₂ _…_j _1m _j ₂₁ _j ₂₂ _…_j _2m _j ₃₁ _j ₃₂ _…_j _3m wherein h and w are eachHeight and width of the target image, delta ₁ 、δ ₂ M is a preset parameter, and delta is satisfied ₁ ＞δ ₂ ，m∈[2,255]，j _i1 、j _i2 、…j _im The gray scale ratio of the pixel values of the ith channel of the target image is more than or equal to delta ₁ Delta for the first m pixel values ₂ For j _i1 、j _i2 、…j _im The ith channel includes one of three red, green and blue RGB channels.

6. The method of any one of claims 1, 4, 5, wherein determining the second characteristic of the target image comprises:

the following character string f will be described _k Determining a second feature of the target image:

f _k ＝l ₁₁ _l ₁₂ _…_l _1m _l ₂₁ _l ₂₂ _…_l _2m _l ₃₁ _l ₃₂ _…_l _3m

wherein l _i1 、l _i2 、…、l _im A gray scale duty ratio of delta or more among a plurality of pixel values of an ith channel of the first image ₁ The ith channel comprises one of three red, green, blue, RGB channels, the first image is an image of a kth portion of the target image, k=1, 2..

Or->

The width of the first image +.>

Or (b)

Or->

Or->

h and w are the height and width of the target image, respectively.

7. The method according to claim 5 or 6, further comprising:

determining the following vector as a second feature of the target image:

u＝(u _i1 ，u _i2 ，…，u _im )，

wherein u is _i1 、u _i2 、…、u _im The pixel values in the ith channel of the target image are j respectively _i1 、j _i2 、…j _im Is used for the gray scale ratio of (a),

for a plurality of pixel values in an ith channel of the target image, a pixel value equal to j _im Is a number of (3).

8. An apparatus for similar image retrieval, the apparatus comprising:

a first determining module, configured to determine a first feature of a target image and a second feature of the target image, where the first feature is used to describe an overall feature of the image, and the second feature is used to describe a local feature of the image;

the screening module is used for determining at least one image to be selected from an image database, wherein the first characteristic of the image to be selected is the same as the first characteristic of the target image;

and the second determining module is used for determining a second image similar to the target image from the at least one image to be selected according to the second characteristic of the target image and the second characteristic of the at least one image to be selected.

9. An electronic device comprising a processor, a memory and a program or instruction stored on the memory and executable on the processor, which program or instruction when executed by the processor implements the steps of the method of any of claims 1-8.

10. A readable storage medium, characterized in that it stores thereon a program or instructions, which when executed by a processor, implement the steps of the method according to any of claims 1-8.

Images