WO2020192262A1

WO2020192262A1 - Physical object image generation method and apparatus, and device

Info

Publication number: WO2020192262A1
Application number: PCT/CN2020/073056
Authority: WO
Inventors: 侯晓楠; 邱雪涛; 万四爽
Original assignee: 中国银联股份有限公司
Priority date: 2019-03-25
Filing date: 2020-01-19
Publication date: 2020-10-01
Also published as: CN109961488A; TWI801717B; TW202105320A

Abstract

The present application discloses a physical object image generation method and apparatus, and a device. The method comprises: performing intrinsic decomposition on a first physical object image to acquire a first reflection image and a first irradiation image of the first physical object image; performing at least one transformation on at least one of the first reflection image and the first irradiation image; and generating at least one second physical object image according to the at least one transformed image, the first reflection image and the first irradiation image.

Description

Method, device and equipment for generating physical image

Cross references to related applications

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office, the application number is 201910227393.2, and the application name is "a method and device for generating physical images" on March 25, 2019, the entire content of which is incorporated into this application by reference in.

Technical field

The present invention relates to the field of computer vision, and in particular to a method, device and equipment for generating a physical image.

Background technique

Image recognition is a technology that uses computers to process, analyze, and understand images to identify targets and objects in various modes. For example, to recognize faces and verify identity. The image recognition model needs to be trained with a large number of real pictures.

For an image recognition model, the more physical pictures, the richer the scenes covered by the training data, and the more accurate the physical recognition. However, the manual collection of physical pictures has great limitations. It is impossible to obtain physical pictures of a part of the scene by adjusting the shooting conditions, especially the subtle changes in shooting conditions. Manually adjusting the shooting conditions will cause large errors in the collection of physical pictures. In some scenes, the physical pictures are missing and the training data is incomplete, which causes the physical recognition model to be unable to accurately identify the physical objects in the missing part of the scene.

Therefore, in the prior art, the manual collection of physical images causes missing physical images in some scenes, and the problem of incomplete training data needs to be solved urgently.

Summary of the invention

The embodiments of the present application provide a method, device, and equipment for generating a physical image, which solves the problem of missing physical images in some scenes and incomplete training data caused by manual collection of physical images in the prior art.

In the first aspect, an embodiment of the present application provides a method for generating a physical image, including:

Perform intrinsic decomposition of the first physical image to obtain a first reflection image and a first illumination image of the first physical image; perform at least one transformation on at least one of the first reflection image and the first illumination image ; According to the transformed at least one image, the first reflected image and the first illuminated image, at least one second physical image is generated.

Optionally, the performing at least one transformation of at least one of the first reflection image and the first illumination image includes: according to M first pixel value transformation rules in a preset illumination transformation algorithm, The pixel values in the first irradiated image undergo M different transformations to obtain different M transformed second irradiated images; wherein, each of the M different transformations is different from the second after the M transformations. One of the illuminated images uniquely corresponds to a second illuminated image; M is a positive integer; and, the at least one second physical image is generated based on the transformed at least one image, the first reflection image, and the first illuminated image , Including: generating M second physical images that are different from the first physical image according to the first reflected image and the M transformed second illuminated images.

Optionally, the performing at least one transformation of at least one of the first reflection image and the first illumination image includes: according to N second pixel value transformation rules in a preset reflection transformation algorithm, The pixel values in the first reflection image undergo N different transformations to obtain different N transformed second reflection images; wherein, each of the N different transformations is different from the first N transformation. One of the two reflection images uniquely corresponds to a second reflection image; N is a positive integer; and the at least one second physical image is generated according to the transformed at least one image, the first reflection image, and the first illumination image , Including: generating N second physical images that are different from the first physical image according to the first illuminated image and the N transformed second reflection images.

Optionally, the performing at least one transformation of at least one of the first reflection image and the first illumination image includes: performing P third pixel value transformation rules in a preset illumination transformation algorithm, The pixel values in the first irradiated image undergo P different transformations to obtain different P transformed third irradiated images; wherein, each transformation in the P different transformations is the same as the P transformation. One of the three illuminated images uniquely corresponds to a third illuminated image; P is a positive integer; according to the Q fourth pixel value conversion rules in the preset reflection conversion algorithm, the pixel values in the first reflection image are different Q times Transformation to obtain Q different transformed third reflection images; wherein each transformation in the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer; and the generating at least one second physical image according to the transformed at least one image, the first reflection image, and the first illuminated image includes: according to the P transformed third images The illumination image and the Q transformed third reflection images are used to generate P*Q second physical images that are different from the first physical image.

Optionally, after the generating at least one second physical image, the method further includes: inputting the at least one second physical image as training data to the image recognition model; according to each second physical image in the training data And update the parameters of the image recognition model with the output result of the second physical image input to the image recognition model.

In the above method, by performing intrinsic decomposition of the first physical image, the first reflection image and the first illumination image of the first physical image are acquired, and then at least one of the first reflection image and the first illumination image is One image undergoes at least one transformation, so at least one illumination image and reflection image after transformation can be obtained, and then at least one image after transformation, the first reflection image and the first illumination image are combined with each other, so that the The initial first physical image generates at least one physical image, and so on, the above steps are performed on each physical image collected manually, which can greatly improve the physical image, make up for the lack of manual collection of physical images in some scenes, and achieve training The effect of data supplementation.

In the second aspect, an embodiment of the present application provides a physical image generation device, including:

The acquisition module is used to perform intrinsic decomposition of the first physical image to acquire the first reflection image and the first illumination image of the first physical image; the processing module is used to perform the analysis of the first reflection image and the first At least one image in the illuminated image is transformed at least once; and used to generate at least one second physical image based on the transformed at least one image, the first reflection image, and the first illuminated image.

Optionally, the processing module is specifically configured to: perform M different transformations on the pixel values in the first illuminated image according to the M first pixel value transformation rules in the preset illumination transformation algorithm to obtain different M transformed second illuminated images; wherein each of the M different transformations uniquely corresponds to one of the M transformed second illuminated images; M is a positive integer; according to The first reflected image and the M transformed second illuminated images generate M second physical images that are different from the first physical image.

Optionally, the processing module is specifically configured to: perform N different transformations on the pixel values in the first reflection image according to the N second pixel value transformation rules in the preset reflection transformation algorithm to obtain different N transformed second reflection images; wherein each of the N different transformations uniquely corresponds to one of the N transformed second reflection images; N is a positive integer; According to the first illuminated image and the N transformed second reflection images, N second physical images that are different from the first physical image are generated.

Optionally, the processing module is specifically configured to: perform P different transformations on the pixel values in the first illuminated image according to the P third pixel value transformation rules in the preset illumination transformation algorithm to obtain different P transformed third illuminated images; wherein each of the P different transformations uniquely corresponds to one of the P transformed third illuminated images; P is a positive integer; According to the Q fourth pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed Q times to obtain different Q transformed third reflection images; where Each of the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer; according to the P transformed third illumination images and The Q transformed third reflection images generate P*Q second physical images that are different from the first physical image.

Optionally, the processing module is further configured to: use the at least one second physical image as training data and input it to the image recognition model; according to each second physical image in the training data, the second physical image is The output result after the physical image is input to the image recognition model, and the parameters of the image recognition model are updated.

In a third aspect, an embodiment of the present application provides a physical image generation device, including:

At least one processor; and,

A memory communicatively connected with the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the physical image generation method described in the first aspect above .

In a fourth aspect, an embodiment of the present application provides a non-transitory computer-readable storage medium that stores computer instructions, and the computer instructions are used to make the computer execute the above-mentioned first aspect. The physical image generation method described.

In a fifth aspect, the embodiments of the present application provide a computer program product, the computer program product includes a calculation program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions are When executed by a computer, the computer is caused to execute the physical image generation method described in the first aspect.

Description of the drawings

FIG. 1 is an overall flowchart of a method for generating a physical image according to Embodiment 1 of the application;

2 is a flow chart of the steps of a method for generating a physical image according to Embodiment 1 of the application;

FIG. 3 is a schematic diagram of the intrinsic decomposition corresponding to a method for generating a physical image according to Embodiment 1 of the application;

4 is a flow chart of the steps of a method for generating a physical image provided in the second embodiment of the application;

FIG. 5 is a flow chart of the steps of a method for generating a physical image according to Embodiment 3 of the application;

FIG. 6 is a schematic structural diagram of a physical image generating device applied to Embodiments 1, 2 and 3 of this application;

FIG. 7 is a schematic structural diagram of a physical image generating device applied to Embodiments 1, 2 and 3 of this application.

detailed description

In order to better understand the above technical solutions, the above technical solutions will be described in detail below with reference to the drawings and specific implementations of the specification. It should be understood that the embodiments of the application and the specific features in the embodiments are detailed to the technical solutions of the application. Note, rather than limiting the technical solution of the present application, the embodiments of the present application and the technical features in the embodiments can be combined with each other if there is no conflict.

Image recognition is widely used in the field of computer vision, such as various face verification systems, which can obtain the authority of the identity by identifying the face of the face, and then perform operations. The tool to realize the function of image recognition is the image recognition model.

Before realizing the image recognition function, the image recognition model needs to be trained with a large number of real pictures. For an image recognition model, the more physical pictures, the richer the scenes covered by the training data, and the more accurate the physical recognition. However, the manual collection of physical pictures has great limitations. It is impossible to obtain physical pictures of a part of the scene by adjusting the shooting conditions, especially the subtle changes in shooting conditions. Manually adjusting the shooting conditions will cause large errors in the collection of physical pictures. In some scenes, the physical pictures are missing and the training data is incomplete, which causes the physical recognition model to be unable to accurately identify the physical objects in the missing part of the scene.

Therefore, the embodiment of the present application proposes a method for generating more physical images based on a small number of manually collected physical images. As shown in FIG. 1, it is an overall flowchart of a method for generating a physical image provided in an embodiment of this application. It should be noted that FIG. 1 only uses a manually collected physical image as an example to illustrate the process, and the manually collected physical image is also collected in multiple scenes.

Given a physical image, use the intrinsic decomposition algorithm to decompose the physical image into a illuminating image (Shading image) and a reflection image (Reflectance image). Among them, the physical image is the initial image collected manually; the illuminated image is the image that reflects the lighting conditions of the original image; the reflected image refers to the image that can remain unchanged under changing lighting conditions, reflecting the texture and material of the original physical image. Among them, the physical image, the illuminated image, and the reflected image are composed of multiple pixels, each pixel has a pixel value, and each pixel is combined to form an image to produce a visual effect. Each pixel has a corresponding pixel value in the physical image, illuminated image, and reflected image, and each pixel value in the physical image, illuminated image, and reflected image corresponds to each other.

Take a physical image collected manually as an example. After obtaining an irradiated image and a reflection image after intrinsic decomposition, the reflection image and the irradiated image are transformed multiple times. After each transformation, an image with the original is obtained. The reflected image or the reflected image or the illuminated image with different pixel values of the illuminated image can be used to generate a large number of physical images different from the original image set by using these modified illuminated images and reflected images. Among them, the brightness of the physical image collected manually is generally affected by the ambient light, and the material information of the physical object itself has nothing to do with the light condition. Therefore, the embodiment of the present application transforms the reflected image into different lighting conditions to obtain multiple transformed reflected images with different lighting conditions; and by transforming the illuminated images with different textures and other conditions, multiple transformed illuminated images with the same lighting condition are obtained. image. It should be noted that the above-mentioned specific methods for transforming the reflected image or the illuminated image are all implemented by transforming the pixel values of the pixels in the reflected image or the illuminated image through a preset algorithm.

Next, in conjunction with Figure 2, the above-mentioned physical image generation method will be further introduced in detail.

As shown in FIG. 2, it is a flowchart of steps of a method for generating a physical image provided in an embodiment of this application.

Step 201: Perform intrinsic decomposition on the first physical image, and obtain a first reflection image and a first illumination image of the first physical image.

In step 201, the relationship among the first physical image (I), the first reflection image (R), and the first illumination image (S) can be expressed by a formula:

I(x,y)=S(x,y)*R(x,y) (1)

Among them (x, y) are the pixel coordinates of the pixel in the image. According to this formula, the reflection map (R) and the illumination map (S) are numerically solved. The schematic diagram is shown in FIG. 3. FIG. 3 is a schematic diagram of the first physical image generation method provided in an embodiment of the application corresponding to the intrinsic decomposition. . The first line is the first physical image, the middle line is the first reflection image, and the last line is the first illumination image. It should be noted that since only the pixel value of the pixel coordinate in the first physical image I is a known quantity, and the pixel value of the first illuminated image corresponding to the pixel coordinate and the pixel value of the first reflection image have more than one set of solutions, therefore In the intrinsic decomposition process, the pixel values of the first illuminated image and the pixel values of the first reflected image that are decomposed are a set of solutions selected at random.

Step 202: Perform at least one transformation on at least one of the first reflected image and the first illuminated image.

Step 203: Generate at least one second physical image according to the transformed at least one image, the first reflection image and the first illumination image.

In step 202, there are three situations:

In the first case, at least one transformation is taken as an example of M transformations. According to the M first pixel value transformation rules in the preset lighting transformation algorithm, the pixel values in the first illuminated image are subjected to M different transformations to obtain different M Two transformed second illuminated images; where each of the M different transformations uniquely corresponds to one of the M transformed second illuminated images; M is a positive integer.

In the second case, according to the N second pixel value transformation rules in the preset reflection transformation algorithm, N different transformations are performed on the pixel values in the first reflection image to obtain different N transformed second reflections Image; wherein, each of the N different transformations is uniquely corresponding to one of the N transformed second reflection images; N is a positive integer.

In the third case, in accordance with the P third pixel value transformation rules in the preset illumination transformation algorithm, P different transformations are performed on the pixel values in the first illumination image to obtain different P transformed third illuminations Image; wherein, each of the P different transformations is uniquely corresponding to one of the P transformed third illumination images; P is a positive integer.

In addition, according to the Q fourth pixel value transformation rules in the preset reflection transformation algorithm, Q different transformations are performed on the pixel values in the first reflection image to obtain Q different transformed third reflection images; wherein Each of the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer.

In the above three cases, take the first case as an example, in which the preset illumination transformation algorithm is encapsulated in an image processing software, such as openCV. When the image processing software calls the lighting condition transformation algorithm, there are multiple first pixel value transformation rules corresponding to multiple lighting condition transformations, that is, one lighting condition corresponds to one first pixel value transformation rule. The pixel value of the first illumination image is transformed according to a first pixel value transformation rule to obtain the transformed illumination image under corresponding lighting conditions. In the second case and the third case, the pixel value conversion rules are also preset according to the lighting conditions or textures. By changing the pixel values, the reflected images and illuminated images that have changed the lighting conditions or textures are obtained, which will not be repeated.

Step 203 corresponds to the situation in step 202, including the following three situations:

In the first case, after the first case in step 202 is performed, according to the first reflected image and the M transformed second illumination images, M second images that are different from the first physical image are generated. Physical image.

In the second case, after the second case in step 202 is performed, according to the first illumination image and the N transformed second reflection images, N second images different from the first physical image are generated. Physical image.

In the third case, after the third case in step 202 is performed, according to the P transformed third illumination images and the Q transformed third reflection images, generate images that are different from the first physical image P*Q second physical image.

To sum up, there are three situations in which the physical image generated in step 202 to step 203 is as follows, which are expressed by the formula as follows:

In the first case, the first reflection image (R) obtained in the intrinsic decomposition stage is kept unchanged, and the first illumination image (S) is modified differently to obtain M different second illumination images (AS ⁽ⁱ⁾ ) , And then use the formula (1) to calculate the generated physical image (AI ⁽ⁱ⁾ ): AI ⁽ⁱ⁾ (x,y)=AS ⁽ⁱ⁾ (x,y)*AR(x,y),i=1, …, M(2) In this case, by modifying the lighting conditions of the first illuminated image, a physical image set A=[AI ⁽¹⁾ ,AI ⁽²⁾ ,...,AI ^(M) ] is generated.

In the second case, keeping the first illumination image (S) obtained in the intrinsic decomposition stage unchanged, make N different modifications to the first reflection image (R) to obtain different reflection images (BR _j ), and then use the formula (1) Calculate the generated picture (BI _j ): BI _j (x,y)=BS(x,y)*BR _j (x,y),j=1,...,N; (3) In this case, By modifying the texture of the first reflection image, a physical image set B=[BI ₁ ,BI ₂ ,...,BI _N ] is generated.

In the third case, the first reflection image (R) is modified Q times differently to obtain a different third reflection image (CR _j ), and each third reflection image (CR _j ) remains unchanged. The irradiation image (S) is modified P times differently to obtain a different third irradiation image (CS _j ), and then the generated image (CI _j ) is calculated using formula (1):

Get generated data set

After step 203, another optional implementation manner is to use the at least one second physical image as training data and input it to the image recognition model; according to each second physical image in the training data, it is related to the first Second, the output result after the physical image is input to the image recognition model, and the parameters of the image recognition model are updated. The generated second physical image greatly increases the amount of training data, which can make the image recognition model more accurate.

As shown in FIG. 4, this is a flow chart of the steps of a method for generating a physical image provided in the second embodiment of the present application. The second embodiment of the present application is a method for generating a multi-illumination face image based on intrinsic decomposition. Illumination changes are the most critical factor that affects the performance of face recognition. The degree of solution to this problem is related to the success or failure of the practical process of face recognition. In order to improve the robustness of the face recognition model to lighting, one of the most direct methods is to add face images under different lighting conditions to the training data. The specific steps are as follows:

Before step 401, collect a physical image set E obtained by manual shooting, for example, E contains 100,000 face images.

Step 401: Perform intrinsic decomposition of each physical image in the physical image set E.

In step 401, for example, k=1, 2...100000, perform intrinsic decomposition on each picture EI _k in the physical image set E to obtain the corresponding reflection image (ER _k ) and illumination image (ES _k ) .

Step 402: Keep the reflected image (ER _k ) unchanged, and perform n different modifications to the illuminated image (ES _k ) according to a preset lighting condition modification algorithm. Wherein, n is an integer greater than 1.

In step 402, each illuminated image ES _k obtains a transformed illuminated image set

Step 403: Generate a physical image set according to the transformed illuminated image set and the reflected image.

Then use the following formula to generate a collection of physical images

Step 404: Determine whether there are any physical images in the data set E for which steps 402 and 403 have not been performed.

If yes, go to step 402; otherwise, use the set of physical images generated from each physical image in E as the final generated training data set E _g . Take the physical image containing 100,000 pictures as an example, E _g =[E ₁ ,…,E ₁₀₀₀₀₀ ], containing 1 million pictures in total; use the data set [E,E _g ] to train the physical object recognition model to obtain the illumination Real object recognition model with more robust conditions.

FIG. 5 is a flow chart of the steps of a method for generating a physical image provided in Embodiment 3 of the present application. Embodiment 3 of the present application is a method for generating training data for image segmentation based on intrinsic decomposition. The purpose of image segmentation is to divide an image into regions with features and extract the target of interest. These features can be pixels, colors, textures, etc. The extraction target can be a single or multiple regions. Specific steps are as follows:

Before step 501, collect a physical image set F obtained by manual shooting, for example, F contains 1000 landscape images.

Step 501: Perform intrinsic decomposition on each physical image in the physical image set F.

In step 501, for example, m=1, 2...1000, perform intrinsic decomposition on each picture FI _m in the physical image set F to obtain the corresponding reflection image (FR _m ) and illumination image (FS _m ) .

Step 502: Keep the reflected image (FR _m ) unchanged, and perform t different modifications to the illuminated image (FS _m ) according to the preset lighting condition modification algorithm. It should be noted that the preset lighting condition transformation algorithm includes multiple pixel value transformation rules, and each pixel value transformation rule corresponds to a transformed reflection image.

In step 502, the transformed illuminated image set is obtained

Step 503: Generate a physical image set according to the transformed irradiation image set

Step 503 uses the following formula for transformation:

Step 504: Keep the illuminated image (FS _m ) unchanged, and perform r different modifications to the reflected image (FR _m ) according to the preset texture modification algorithm. It should be noted that the preset texture transformation algorithm includes multiple pixel value transformation rules, and each pixel value transformation rule corresponds to a transformed reflection image.

In step 504, the transformed reflection image set [FR _m,1 ,...,FR _m,r ] is obtained.

Step 505: Generate a physical image set F _m ′=[FI _m,1 ,...,FI _m,r ] according to the transformed reflection image set.

In step 505, the following formula is used for transformation:

FI _m,j (x,y)=FS _m (x,y)*FR _m,j (x,y), j=1,...,r.

Step 506: Determine whether there are any physical images in the data set F for which step 502 and step 504 are not performed. If yes, go to step 502; otherwise, use the set of physical images generated from each physical image in E as the final generated training data set F _h =[F ₁ , F ₁ ′..., F _t , F _r ′]. For example, when t=r=10, F _h contains a total of 20,000 physical images. Using the data set [F, F _h ] to train the image segmentation model, and using the data set with richer lighting conditions, colors and material conditions for model training, will greatly improve its accuracy.

In the first, second and third embodiments of this application, a method of generating a large amount of training data based on a small number of manually-collected physical pictures, by intrinsically decomposing the manually-collected physical pictures to obtain the illumination map and the reflection map, According to the changes in lighting conditions or textures, the reflection map is modified to modify the illumination map and the reflection map to generate a more abundant type of lighting and texture. The training data of the image recognition model is expanded, so that the image recognition model can be used for real objects in different scenes. The recognition is more accurate and robust.

The embodiment of the present application greatly reduces the manpower input; through the above method, a large number of physical images can be generated as training data, thereby greatly reducing the cost of data accumulation, and collecting a large amount of training data in a shorter time. In addition, a large amount of training data can be customized according to specific application scenarios; this method can generate a training data set containing more abundant lighting and texture types on the basis of the original physical image set; the generated data can train A model that is more robust to lighting effects and more general to different scenes can improve the performance of computer vision fields such as face or object detection and recognition, and image segmentation.

As shown in FIG. 6, it is a schematic structural diagram of a physical image generating device applied to Embodiments 1, 2 and 3 of this application.

The embodiment of the present application provides a physical image generation device, including:

The acquisition module 601 is used to perform intrinsic decomposition of the first physical image to acquire the first reflection image and the first illumination image of the first physical image; the processing module 602 is used to perform the intrinsic decomposition of the first physical image and the At least one image in the first illumination image is transformed at least once; and for generating at least one second physical image based on the transformed at least one image, the first reflection image, and the first illumination image.

Optionally, the processing module 602 is specifically configured to: perform M different transformations on the pixel values in the first illuminated image according to the M first pixel value transformation rules in the preset illumination transformation algorithm to obtain different M transformed second illuminated images of M; wherein, each of the M different transformations uniquely corresponds to one of the M transformed second illuminated images; M is a positive integer; According to the first reflected image and the M transformed second illumination images, M second physical images that are different from the first physical image are generated.

Optionally, the processing module 602 is specifically configured to: perform N different transformations on the pixel values in the first reflection image according to the N second pixel value transformation rules in the preset reflection transformation algorithm to obtain different N transformed second reflection images; wherein, each of the N different transformations uniquely corresponds to one of the N transformed second reflection images; N is a positive integer ; According to the first illumination image and the N transformed second reflection images, N second physical images that are different from the first physical image are generated.

Optionally, the processing module 602 is specifically configured to: perform P different transformations on the pixel values in the first illuminated image according to the P third pixel value transformation rules in the preset illumination transformation algorithm to obtain different P transformed third irradiated images; wherein, each of the P different transformations uniquely corresponds to one of the P transformed third irradiated images; P is a positive integer According to the Q fourth pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed Q times to obtain different Q transformed third reflection images; wherein, Each of the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer; according to the P transformed third illumination images And the Q transformed third reflection images to generate P*Q second physical images different from the first physical image.

Optionally, the processing module 602 is further configured to: use the at least one second physical image as training data and input it into the image recognition model; according to each second physical image in the training data, it is related to the first physical image. Second, the output result after the physical image is input to the image recognition model, and the parameters of the image recognition model are updated.

Based on the same technical concept, the embodiment of the present application provides a physical image generating device. At least one processor; and, a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable The at least one processor can execute the physical image generation method in the foregoing embodiment.

Taking a processor as an example, FIG. 7 is a structure of a physical image generating device provided by an embodiment of the application. The physical image generating device 700 includes a transceiver 701, a processor 702, a memory 703, and a bus system 704;

Among them, the memory 703 is used to store programs. Specifically, the program may include program code, and the program code includes computer operation instructions. The memory 703 may be a random access memory (random access memory, RAM for short), or a non-volatile memory (non-volatile memory), such as at least one disk memory. Only one memory is shown in the figure. Of course, the memory can also be set to multiple as required. The memory 703 may also be a memory in the processor 702.

The memory 703 stores the following elements, executable modules or data structures, or their subsets, or their extended sets:

Operating instructions: including various operating instructions, used to implement various operations.

Operating system: including various system programs, used to implement various basic services and process hardware-based tasks.

The foregoing method for generating an object image in the embodiment of the present application may be applied to the processor 702, or implemented by the processor 702. The processor 702 may be an integrated circuit chip with signal processing capabilities. In the implementation process, the steps of the foregoing physical image generation method can be completed by hardware integrated logic circuits in the processor 702 or instructions in the form of software. The above-mentioned processor 702 may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware Components. The methods, steps, and logical block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed 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, registers. The storage medium is located in the memory 703, and the processor 702 reads the information in the memory 703, and performs the following steps in combination with its hardware:

The transceiver 701 is configured to perform intrinsic decomposition of the first physical image to obtain a first reflection image and a first illumination image of the first physical image;

The processor 702 is configured to perform at least one transformation on at least one of the first reflected image and the first illuminated image; and configured to perform at least one transformation based on the transformed at least one image, the first reflected image, and the The first irradiated image generates at least one second physical image.

Preferably, the processor 702 is specifically configured to:

According to the M first pixel value transformation rules in the preset illumination transformation algorithm, perform M different transformations on the pixel values in the first illumination image to obtain different M transformed second illumination images; wherein, the Each of the M different transformations uniquely corresponds to one of the M transformed second illumination images; M is a positive integer; according to the first reflection image and the M transformations To generate M second physical images that are different from the first physical image.

Preferably, the processor 702 is specifically configured to:

According to the N second pixel value transformation rules in the preset reflection transformation algorithm, perform N different transformations on the pixel values in the first reflection image to obtain different N transformed second reflection images; Each of the N different transformations uniquely corresponds to a second reflection image among the N transformed second reflection images; N is a positive integer; according to the first illumination image and the N transformations After the second reflection image, N second physical images different from the first physical image are generated.

Preferably, the processor 702 is specifically configured to:

According to the P third pixel value transformation rules in the preset illumination transformation algorithm, the pixel values in the first illumination image are transformed P different times to obtain different P transformed third illumination images; Each of the P different transformations uniquely corresponds to one of the P transformed third illumination images; P is a positive integer; according to the Q fourth pixel values in the preset reflection transformation algorithm The transformation rule is to perform Q different transformations on the pixel values in the first reflection image to obtain different Q transformed third reflection images; wherein, each of the Q different transformations is different from the One of the Q transformed third illumination images uniquely corresponds to one third illumination image; Q is a positive integer; according to the P transformed third illumination images and the Q transformed third reflection images, generate P*Q second physical images different from the first physical image.

Preferably, the processor 702 is further configured to:

Use the at least one second physical image as training data and input it to the image recognition model; according to each second physical image in the training data and the output result of the second physical image input to the image recognition model To update the parameters of the image recognition model.

The physical image generation equipment in the embodiments of this application exists in various forms, including but not limited to:

(1) Ultra-mobile personal computer equipment: This type of equipment belongs to the category of personal computers, has calculation and processing functions, and generally also has mobile Internet features. Such terminals include: PDA, MID and UMPC devices, such as iPad.

(2) Other electronic devices with physical image generation function.

Those skilled in the art can understand that all or part of the steps in the method of the above-mentioned embodiments can be implemented by instructing relevant hardware through a program. The program is stored in a storage medium and includes several instructions to enable a device (which can be a single-chip microcomputer). , A chip, etc.) or a processor (processor) executes all or part of the steps of the method in each embodiment of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program code .

In addition, the present application also provides a non-transitory computer-readable storage medium that stores computer instructions, and the computer instructions are used to make the computer execute any of the above-mentioned physical objects. Image generation method.

In addition, the present application also provides a computer program product, the computer program product includes a calculation program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer , Enabling the computer to execute any of the above-mentioned physical image generation methods.

Finally, it should be noted that those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program codes.

This application is described with reference to flowcharts and/or block diagrams of methods, equipment (systems), and computer program products according to this application. It should be understood that each process and/or block in the flowchart and/or block diagram, and the combination of processes and/or blocks in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing equipment to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing equipment are generated It is a device that realizes the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device. The device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

Obviously, those skilled in the art can make various changes and modifications to the application without departing from the scope of the application. In this way, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalent technologies, this application also intends to include these modifications and variations.

Claims

A method for generating a physical image, which is characterized in that it includes:

Performing intrinsic decomposition on the first physical image to obtain a first reflection image and a first illuminated image of the first physical image;

Performing at least one transformation on at least one of the first reflected image and the first illuminated image;

According to the transformed at least one image, the first reflection image and the first illumination image, at least one second physical image is generated.
The method according to claim 1, wherein said transforming at least one of said first reflected image and said first illuminated image at least once comprises:

According to the M first pixel value transformation rules in the preset illumination transformation algorithm, perform M different transformations on the pixel values in the first illumination image to obtain different M transformed second illumination images; wherein, the Each of the M different transformations uniquely corresponds to one of the M transformed second illumination images; M is a positive integer;

The generating at least one second physical image according to the transformed at least one image, the first reflection image, and the first illumination image includes:

According to the first reflected image and the M transformed second illumination images, M second physical images that are different from the first physical image are generated.
The method according to claim 1, wherein said transforming at least one of said first reflected image and said first illuminated image at least once comprises:

According to the N second pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed N times to obtain different N transformed second reflection images; Each of the N times of different transformations uniquely corresponds to one of the N transformed second reflection images; N is a positive integer;

The generating at least one second physical image according to the transformed at least one image, the first reflection image, and the first illumination image includes:

According to the first illuminated image and the N transformed second reflection images, N second physical images that are different from the first physical image are generated.
The method according to claim 1, wherein said transforming at least one of said first reflected image and said first illuminated image at least once comprises:

According to the P third pixel value transformation rules in the preset illumination transformation algorithm, the pixel values in the first illumination image are transformed P different times to obtain different P transformed third illumination images; Each of the P different transformations uniquely corresponds to one of the P transformed third illumination images; P is a positive integer;

According to the Q fourth pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed Q times to obtain different Q transformed third reflection images; where Each of the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer;

The generating at least one second physical image according to the transformed at least one image, the first reflection image, and the first illumination image includes:

According to the P transformed third illumination images and the Q transformed third reflection images, P*Q second physical images that are different from the first physical image are generated.
The method according to any one of claims 1 to 4, wherein after said generating at least one second physical image, the method further comprises:

Input the at least one second physical image as training data to the image recognition model;

According to each second physical image in the training data and the output result after the second physical image is input to the image recognition model, the parameters of the image recognition model are updated.
A physical image generating device, characterized in that it comprises:

An acquisition module, configured to perform intrinsic decomposition of the first physical image, and acquire the first reflection image and the first illumination image of the first physical image;

A processing module, configured to perform at least one transformation on at least one of the first reflected image and the first illuminated image;

And it is used to generate at least one second physical image according to the transformed at least one image, the first reflection image and the first illumination image.
The device according to claim 6, wherein the processing module is specifically configured to:

According to the M first pixel value transformation rules in the preset illumination transformation algorithm, perform M different transformations on the pixel values in the first illumination image to obtain different M transformed second illumination images; wherein, the Each of the M different transformations uniquely corresponds to one of the M transformed second illumination images; M is a positive integer;

According to the first reflected image and the M transformed second illumination images, M second physical images that are different from the first physical image are generated.
The device according to claim 6, wherein the processing module is specifically configured to:

According to the N second pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed N times to obtain different N transformed second reflection images; Each of the N times of different transformations uniquely corresponds to one of the N transformed second reflection images; N is a positive integer;

According to the first illuminated image and the N transformed second reflection images, N second physical images that are different from the first physical image are generated.
The device according to claim 6, wherein the processing module is specifically configured to:

According to the P third pixel value transformation rules in the preset illumination transformation algorithm, the pixel values in the first illumination image are transformed P different times to obtain different P transformed third illumination images; Each of the P different transformations uniquely corresponds to one of the P transformed third illumination images; P is a positive integer;

According to the Q fourth pixel value transformation rules in the preset reflection transformation algorithm, the pixel values in the first reflection image are transformed Q times to obtain different Q transformed third reflection images; where Each of the Q different transformations uniquely corresponds to one of the Q transformed third illumination images; Q is a positive integer;

According to the P transformed third illumination images and the Q transformed third reflection images, P*Q second physical images different from the first physical image are generated.
The device according to any one of claims 6-9, wherein the processing module is also used for:

Input the at least one second physical image as training data to the image recognition model;

According to each second physical image in the training data and the output result after the second physical image is input to the image recognition model, the parameters of the image recognition model are updated.
A physical image generating device, characterized in that it comprises:

At least one processor; and,

A memory communicatively connected with the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the physical image of any one of claims 1-5 Generation method.
A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute the physical object described in any one of claims 1-5 Image generation method.
A computer program product, characterized in that, the computer program product includes a calculation program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed by a computer, The computer executes the physical image generation method of any one of claims 1-5.