CN111951176A - A method and device for removing partial shadows from face images - Google Patents

Abstract

The invention discloses a method and device for removing partial shadows of a face image. The method obtains a to-be-processed area including a shadow area based on a preset face photo, the preset face photo includes the to-be-processed face image, and the The face image to be processed includes the area to be processed, the shadow area is extracted from the area to be processed according to a preset pixel threshold, the shadow area is gamma corrected based on a preset gamma coefficient, and the shadow area is gamma corrected based on the preset gamma coefficient. The result of the gamma correction obtains a result area, and generates a face photo with partial shadows removed according to the result area, thereby reducing the processing traces when removing the partial shadows of the face image, thereby improving the quality of the 3D face model.

Description

A method and device for removing partial shadows from face images

technical field

The present application relates to the field of image processing, and more particularly, to a method and device for removing partial shadows from a face image.

Background technique

In the process of scanning a face with a camera to form a 3D face reconstruction, it is necessary to remove the illumination information as much as possible to collect the pixel information in the face photo. However, the illumination information of some local areas cannot be removed by the global de-illumination method. For example, the shadow generated by the nostrils in the photo is often ignored by the global de-illumination algorithm, and the purpose of removing the shadow cannot be achieved. These shadows will cause the 3D face model to appear redundant shadow information in the 3D scene. As shown in Figure 1, the part in the circle is the effect of not processing the shadow of the nostril position.

In order to remove shadows in local areas, in the prior art, an algorithm of target recognition is used to find the shadows at local locations, and then the pixels in the shadows are replaced by counting the composition of pixels in the image. The new image generated in this way has obvious traces in the original shadow position, and the edges are also particularly prominent, giving people a bad visual perception. As shown in the circled part in Figure 2, the altered marks at the nostrils are too obvious.

Also widely studied in academia, Kumar proposes a hypothesis test to detect shadows in images and then remove shadows from images using the energy function concept. In this method, the shadow detection time is longer, and the traces of shadow processing are obvious. Kaushik Deb proposes a simple framework using the luminance-chroma:blue, chroma:red (YCbCr) color space for detecting and eliminating shadows in images. Although this method improves the detection and processing speed, the final effect of removing shadows is still not ideal, as shown in Figure 3.

To sum up, when the prior art removes the partial shadow of the face image, since the contour information of the local area and the skin pixel information still existing in the shadow cannot be preserved, obvious processing marks are caused and the quality of the 3D face model is reduced.

Therefore, how to reduce the processing traces when removing the partial shadows of the face image, thereby improving the quality of the 3D face model, is a technical problem to be solved at present.

SUMMARY OF THE INVENTION

Due to the technical problem that obvious traces will appear when removing partial shadows from face images in the prior art, the present invention provides a method for removing partial shadows from face images, the method comprising:

Acquiring a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area;

Extracting the shadow area from the to-be-processed area according to a preset pixel threshold;

Perform gamma correction on the shadow area based on a preset gamma coefficient, and obtain a result area based on a result of the gamma correction;

A face photo with partial shadows removed is generated according to the result area.

Preferably, the to-be-processed area including the shadow area is obtained based on the preset face photo, specifically:

extracting the to-be-processed face image from the preset face photo based on the first preset cascade classifier;

The to-be-processed region is extracted from the to-be-processed face image based on a second preset cascaded classifier.

Preferably, the to-be-processed region is extracted from the to-be-processed face image based on a second preset cascade classifier, specifically:

Obtaining a plurality of recognition results corresponding to the to-be-processed area from the to-be-processed face image based on the second preset cascade classifier;

The to-be-processed area is extracted from a plurality of the identification results according to a preset verification rule, and the preset verification rule is determined according to a preset positional relationship of the facial features or a preset proportional relationship of the facial features.

Preferably, the shadow area is extracted from the to-be-processed area according to a preset pixel threshold, specifically:

A binary image is segmented from the grayscale image of the region to be processed according to the preset pixel threshold, which is determined according to the Otsu method;

The shaded area is extracted from the binary image.

Preferably, gamma correction is performed on the shadow area based on a preset gamma coefficient, specifically:

Iteratively adjusts the number of shaded pixels in the shaded area based on the preset gamma coefficient, until the number of shaded pixels is less than the preset number to determine a result of the gamma correction.

Preferably, a face photo with partial shadows removed is generated according to the result area, specifically:

generating a processed region based on replacing the shadow region with the result region;

generating a processed face image based on replacing the processed region with the to-be-processed region;

The partial shadow-removed face photo is generated based on replacing the processed face image with the to-be-processed face image.

Correspondingly, the present invention also provides a device for removing partial shadows from a face image, the device comprising:

a first obtaining module, configured to obtain a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area;

an extraction module, configured to extract the shadow area from the to-be-processed area according to a preset pixel threshold;

a second obtaining module, configured to perform gamma correction on the shadow area based on a preset gamma coefficient, and obtain a result area based on the result of the gamma correction;

A generating module is configured to generate a face photo with partial shadow removed according to the result area.

Preferably, the first acquisition module is specifically used for:

extracting the to-be-processed face image from the preset face photo based on the first preset cascade classifier;

The to-be-processed region is extracted from the to-be-processed face image based on a second preset cascaded classifier.

Preferably, the first acquisition module is also specifically used for:

Obtaining a plurality of recognition results corresponding to the to-be-processed area from the to-be-processed face image based on the second preset cascade classifier;

The to-be-processed area is extracted from a plurality of the identification results according to a preset verification rule, and the preset verification rule is determined according to a preset positional relationship of the facial features or a preset proportional relationship of the facial features.

Preferably, the extraction module is specifically used for:

A binary image is segmented from the grayscale image of the region to be processed according to the preset pixel threshold, which is determined according to the Otsu method;

The shaded area is extracted from the binary image.

Preferably, the second acquisition module is specifically used for:

Iteratively adjusts the number of shaded pixels in the shaded area based on the preset gamma coefficient, until the number of shaded pixels is less than the preset number to determine a result of the gamma correction.

Preferably, the generation module is specifically used for:

generating a processed region based on replacing the shadow region with the result region;

generating a processed face image based on replacing the processed region with the to-be-processed region;

The partial shadow-removed face photo is generated based on replacing the processed face image with the to-be-processed face image.

Correspondingly, the present application also proposes a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the instructions are executed on a terminal device, the terminal device is caused to perform the above-mentioned removal. Methods for local shading of face images.

Correspondingly, the present application also proposes a computer program product, which, when running on a terminal device, enables the terminal device to execute the above-described method for removing partial shadows from a face image.

Compared with the prior art, the present invention has the following beneficial effects:

The invention discloses a method and a device for removing partial shadows of a face image. A to-be-processed area including a shadow area is obtained based on a preset face photo, the preset face photo includes a to-be-processed face image, and the to-be-processed face image is obtained. The face image includes the area to be processed, the shadow area is extracted from the area to be processed according to a preset pixel threshold, gamma correction is performed on the shadow area based on a preset gamma coefficient, and based on the gamma The result of the correction obtains a result area, and generates a face photo with partial shadows removed according to the result area. By removing the partial shadows of the face image, processing traces when removing the partial shadows of the face image can be reduced, and the 3D face model can be improved. The quality of 3D face reconstruction in high image quality is satisfied.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can also be obtained from these drawings without creative effort.

Fig. 1 shows the 3D human face renderings without removing the nostril shadow in the prior art;

Fig. 2 shows the naive shadow detection and removal effect diagram in the prior art;

Fig. 3 shows the prior art based on YCbCr color space to detect and eliminate the shadow in the image;

4 shows a schematic flowchart of a method for removing partial shadows from a face image provided by an embodiment of the present invention;

5 shows a schematic flowchart of a method for removing partial shadows from a face image proposed in another embodiment of the present invention;

Fig. 6 shows the region diagram of the face and nose obtained by the cascade classifier in the embodiment of the present invention;

FIG. 7 shows a shadow removal effect diagram in an embodiment of the present invention;

Fig. 8 shows the effect diagram after the nostril is removed shadow in the embodiment of the present invention;

FIG. 9 shows a schematic structural diagram of a device for removing partial shadows from a face image according to an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

As described in the background art, when removing the partial shadow of a face image in the prior art, since the contour information of the local area and the skin pixel information still existing in the shadow cannot be preserved, obvious processing traces are caused, and the 3D face model is reduced. the quality of.

In order to solve the above problem, an embodiment of the present invention provides a method for removing partial shadows from a face image, and obtains a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes the to-be-processed face image. , the to-be-processed face image includes the to-be-processed area, extracts the shadow area from the to-be-processed area according to a preset pixel threshold, performs gamma correction on the shadow area based on a preset gamma coefficient, and Obtaining a result area based on the result of the gamma correction, and generating a face photo with partial shadows removed from the result area, reducing the processing traces when removing the partial shadows of the face image, and improving the quality of the 3D face model .

FIG. 4 is a schematic flowchart of a method for removing partial shadows from a face image proposed by an embodiment of the present invention, and the method includes the following steps:

S401. Acquire a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area. In a specific implementation scenario, first obtain the to-be-processed area including the shadow area in the face photo. The face photo can be a to-be-processed face image or other pictures including a human face, and the to-be-processed area can be the nostril area or other areas.

In order to accurately obtain the to-be-processed area including the shadow area, in a preferred embodiment of the present application, the to-be-processed area including the shadow area is obtained based on a preset face photo, specifically:

extracting the to-be-processed face image from the preset face photo based on the first preset cascade classifier;

The to-be-processed region is extracted from the to-be-processed face image based on a second preset cascaded classifier.

In a specific implementation scenario, since the features of the face image are more obvious, and it is more efficient to obtain the area to be processed on the face image, the face image to be processed is first obtained according to the face photo, and then For obtaining the to-be-processed area on the face image, the target image detection methods are mainly divided into two categories, rule-based and statistics-based. Because the rule-based method is more complicated and requires the participation of a large number of experts and professionals, the statistics-based method is chosen. Although the neural network has a very high accuracy, in order to use this function on mobile devices, a smaller model and less CPU/GPU consumption are required. Therefore, cascade classifiers in traditional machine learning are used to identify the location of the target image in the picture. The steps of identifying the target image in the image by the cascade classifier include: extracting image features through Harr, then using the Adaboost algorithm to train a strong classifier, and finally distinguishing between the target image and the non-target image.

In order to facilitate the distinction, in a preferred embodiment of the present application, the face image to be processed is extracted from the preset face photo by the first preset cascade classifier, and the face image to be processed is extracted from the preset face photo by the second preset cascade classifier. The to-be-processed area is extracted from the to-be-processed face image.

It should be noted that the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other methods of obtaining the to-be-processed area including the shadow area based on the preset face photo all belong to the protection scope of the present application.

In order to improve the accuracy of extracting the to-be-processed area, in a preferred embodiment of the present application, the to-be-processed area is extracted from the to-be-processed face image based on a second preset cascade classifier, specifically:

Obtaining a plurality of recognition results corresponding to the to-be-processed area from the to-be-processed face image based on the second preset cascade classifier;

The to-be-processed area is extracted from a plurality of the identification results according to a preset verification rule, and the preset verification rule is determined according to a preset positional relationship of the facial features or a preset proportional relationship of the facial features.

In a specific implementation scenario, since the traditional machine learning algorithm is adopted in this preferred embodiment, the region to be processed is obtained from the face image to be processed through the second preset cascade classifier, and multiple recognition results will be generated, which is When the verification rule needs to be set again, the correct to-be-processed area is extracted from the plurality of identification results according to the set verification rule as the unique result. Since the to-be-processed area is obtained from a face image, in a preferred embodiment of the present application, the preset verification rule is determined according to the preset positional relationship of the facial features or the preset proportional relationship of the facial features . Of course, the to-be-processed area can also be extracted from a plurality of the identification results through other preset verification rules according to the actual situation. Extract the to-be-processed area.

It should be noted that the solution of the above preferred embodiment is only a specific implementation solution proposed in this application, and the other is based on the second preset cascade classifier to extract the to-be-processed area from the to-be-processed face image. All methods belong to the protection scope of the present application.

S402: Extract the shadow area from the to-be-processed area according to a preset pixel threshold.

In a specific implementation scenario, in the acquired to-be-processed area including the shadow area, the pixel values of the shadow area and the non-shadow area are different, and the non-shadow area that is not within the threshold value is excluded by a preset threshold to obtain the result. the shaded area.

In order to accurately extract the shadow area, in a preferred embodiment of the present application, the shadow area is extracted from the to-be-processed area according to a preset pixel threshold, specifically:

A binary image is segmented from the grayscale image of the region to be processed according to the preset pixel threshold, which is determined according to the Otsu method;

The shaded area is extracted from the binary image.

In a specific implementation scenario, a binary image means that each pixel on the image has only two possible values or gray-level states, and the gray-scale image of the to-be-processed area is divided according to the preset pixel threshold Each pixel of is divided into two types, that is, one that meets the threshold, and one that does not meet the threshold, and then the binary image of the grayscale image of the area to be processed can be obtained.

The Otsu method is an algorithm for determining the threshold value of image binarization. After the image binarization segmentation is performed according to the threshold obtained by the Otsu method, the inter-class variance of the foreground and background images is the largest. In order to minimize the probability of misclassification when segmenting each pixel in the grayscale image of the region to be processed, in a preferred embodiment of the present application, the preset pixel threshold is determined according to the Otsu method.

In the binary image segmented from the grayscale image of the area to be processed according to the preset pixel threshold, there is one color for the non-shaded area and one color for the shaded area. At this time, it is easy to extract the binary image from the The shaded area is extracted.

It should be noted that the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other methods of extracting the shadow area from the to-be-processed area according to the preset pixel threshold all belong to the protection of the present application. scope.

S403, perform gamma correction on the shadow area based on a preset gamma coefficient, and obtain a result area based on a result of the gamma correction.

In a specific implementation scenario, gamma correction is used to perform nonlinear operations or inverse operations on the brightness or tristimulus values of light in a film or imaging system, because the human eye's perception of brightness is not related to physical power. is proportional to the power function, so in the preferred embodiment of the present application, a preset gamma coefficient is used to perform gamma correction on the shadow area, after the correction, the brightness of the shadow area can be increased, and then according to the The result of the gamma correction is obtained in the result area.

In order to improve the result of gamma correction, in a preferred embodiment of the present application, gamma correction is performed on the shadow area based on a preset gamma coefficient, specifically:

Iteratively adjusts the number of shaded pixels in the shaded area based on the preset gamma coefficient, until the number of shaded pixels is less than the preset number to determine a result of the gamma correction.

In a specific implementation scenario, in order to gradually increase the brightness of the shadow area, the preset gamma coefficient is generally selected as a value less than 1 and close enough to 1. The horse coefficient iteratively adjusts the number of shadow pixels in the shadow area for several times. When the number of shadow pixels is less than the preset number, the iteration is stopped and the result of this iteration is the final result of the gamma correction. It should be noted that, in order to make the final result infinitely close to ideal, the adjustment range of the preset gamma coefficient in each iteration should not be too large, and the smaller the adjustment range, the better without affecting the correction speed.

It should be noted that the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other methods of performing gamma correction on the shadow area based on preset gamma coefficients all belong to the protection scope of the present application.

S404, generating a face photo with partial shadows removed according to the result area.

In a specific implementation scenario, after the result area is obtained, the shadow area in the face photo is replaced with the result area to obtain a face photo with partial shadow removed.

In order to accurately generate a face photo with partial shadow removed, in a preferred embodiment of the present application, a face photo with partial shadow removed is generated according to the result area, specifically:

generating a processed region based on replacing the shadow region with the result region;

generating a processed face image based on replacing the processed region with the to-be-processed region;

The partial shadow-removed face photo is generated based on replacing the processed face image with the to-be-processed face image.

In a specific implementation scenario, after obtaining the result area, first replace the shadow area with the result area, obtain the processed area corresponding to the to-be-processed area after replacing the shadow area, and then replace the shadow area with the result area. The processed area replaces the to-be-processed area to generate a processed face image, and finally replaces the processed face image with the to-be-processed face image to generate the final partial shadow-removed face photo.

It should be noted that the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other methods of generating face photos with partial shadows removed according to the result area all belong to the protection scope of the present application.

By applying the above technical solutions, a to-be-processed area including a shadow area is obtained based on a preset face photo, the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area, according to Extracting the shadow area from the to-be-processed area with a preset pixel threshold, performing gamma correction on the shadow area based on a preset gamma coefficient, and obtaining a result area based on a result of the gamma correction, and according to the result Regional generation of face photos with partial shadows removed can reduce the processing traces when removing partial shadows of face images, improve the quality of 3D face models, and then meet the needs of 3D face reconstruction in mobile games, mobile APPs and virtual reality. high-definition requirements.

In order to further illustrate the technical idea of the present invention, the technical solutions of the present invention are now described with reference to specific application scenarios.

The embodiment of the present invention proposes a method for generating a three-dimensional face model. The face is recognized in a picture, the position of the nose in the face is obtained, the shadow area is detected, and gamma correction is used to generate a shadow-free photo of the face. The implementation process is as follows: As shown in Figure 5, the specific steps of the above method are as follows:

S501, recognize the face in the image.

In order to accurately recognize the face in the image, in the preferred embodiment of the present application, since the features of the face are more obvious, the accuracy of recognizing the nose in the area of the face is higher, so in the present invention, it is necessary to first identify Face, specifically, using the cascade classifier in traditional machine learning to identify the position of the face in the picture. That is, the image features are extracted by Harr, and then a strong classifier is trained by using the Adaboost algorithm to distinguish between human faces and non-human faces.

S502, obtain the position of the nose in the human face.

In order to accurately locate the position of the nose in the face, in a preferred embodiment of the present application, as shown in FIG. 6 , the same algorithm as in S201 is used to obtain the area of the nose. The red box in Figure 6 is the identified nose area. Due to the use of traditional machine learning algorithms, in order to improve accuracy, it is necessary to manually preset some limited rules. For example, only using machine learning will recognize multiple noses. Therefore, in this application, prior knowledge is added to determine which one is the real nose. , and must conform to the positional relationship, proportional relationship, etc. between the facial features.

It should be noted that, the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other predetermined and limited rules are different within the protection scope of the present application.

S503, the shadow area is detected and corrected by gamma.

In order to more accurately correct the shadow area, in a preferred embodiment of the present application, as shown in Figure 7(A), a threshold is first used to obtain a binary image from a grayscale image, or it is used to remove noise, that is, to filter out too small or too large Then use the Otsu method algorithm to obtain the optimal threshold value, and finally the extracted shadow area is shown in Figure 7(B).

It should be noted that, the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other different ways of obtaining the threshold value all belong to the protection scope of the present application.

In this step, the RGB value and the power are not in a simple linear relationship, but in a power function relationship, so gamma correction is used to adjust the brightness of the shadow area. Gamma correction is used to perform nonlinear operations or inverse operations on the luminance or tristimulus values of light in a film or imaging system. The RGB pixel value of the shadow area is changed by formula (1).

Vout=A×Vin ^γ (1)

After the first iterative calculation, that is, the same method is processed multiple times, as shown in Figure 7(C), on the basis of retaining the skin color, the shadow area becomes smaller. After another iteration, the shadow area becomes smaller, as shown in Figure 7(D). End the iteration until the shadow area is small enough to a certain number of shadow pixels.

S504, generate a face shadow removal photo.

In order to generate a shadow-removed face photo, in a preferred embodiment of the present application, the shadow-removed area is obtained and the original nostril shadowed area is replaced by pixels, the nose area is replaced by pixels, the original nose area is replaced by pixels, and the face area is replaced by pixels Replace the original face area, and finally obtain a face photo with the nostril shadow removed, as shown in the left image of Figure 8. The right picture in Figure 8 is the effect of the 3D face model generated by the photo after de-shading in the 3D scene, and there is no bad situation caused by shadows around the nostrils.

It should be noted that, the solution of the above preferred embodiment is only a specific implementation solution proposed by the present application, and other ways of replacing the original image belong to the protection scope of the present application.

By applying the above technical solutions, the face is recognized in the picture, the position of the nose in the face is obtained, the shadow area is detected and the gamma correction is used to generate a face shadow-removed photo, which reduces the problem of removing the partial shadow of the face image. Processed traces to improve the quality of 3D face models

Corresponding to the method for processing a face image proposed by the embodiment of the present application, the embodiment of the present application also proposes a device for processing a face image. As shown in FIG. 9 , the device includes:

The first obtaining module 901 obtains a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area;

The extraction module 902, extracts the shadow area from the to-be-processed area according to a preset pixel threshold;

The second obtaining module 903 performs gamma correction on the shadow area based on a preset gamma coefficient, and obtains a result area based on the result of the gamma correction;

The generating module 904 generates a face photo with partial shadow removed according to the result area.

In the specific application scenario of the present application, the first acquisition module 901 is specifically used for:

extracting the to-be-processed face image from the preset face photo based on the first preset cascade classifier;

The to-be-processed region is extracted from the to-be-processed face image based on a second preset cascaded classifier.

In the specific application scenario of the present application, the first acquisition module 901 is also specifically used for:

Obtaining a plurality of recognition results corresponding to the to-be-processed area from the to-be-processed face image based on the second preset cascade classifier;

The to-be-processed area is extracted from a plurality of the identification results according to a preset verification rule, and the preset verification rule is determined according to a preset positional relationship of the facial features or a preset proportional relationship of the facial features. In the specific application scenario of the present application, the extraction module 902 is specifically used for:

A binary image is segmented from the grayscale image of the region to be processed according to the preset pixel threshold, which is determined according to the Otsu method;

The shaded area is extracted from the binary image.

In the specific application scenario of the present application, the second acquisition module 903 is specifically used for:

Iteratively adjusts the number of shaded pixels in the shaded area based on the preset gamma coefficient, until the number of shaded pixels is less than the preset number to determine a result of the gamma correction.

In the specific application scenario of the present application, the generating module 904 is specifically used for:

generating a processed region based on replacing the shadow region with the result region;

generating a processed face image based on replacing the processed region with the to-be-processed region;

The partial shadow-removed face photo is generated based on replacing the processed face image with the to-be-processed face image.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or some technical features thereof are equivalently replaced; and these modifications or replacements do not drive the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims (9)

1. a method for removing partial shadow of face image, is characterized in that, described method comprises: Acquiring a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area; Extracting the shadow area from the to-be-processed area according to a preset pixel threshold; Perform gamma correction on the shadow area based on a preset gamma coefficient, and obtain a result area based on a result of the gamma correction; A face photo with partial shadows removed is generated according to the result area. 2. method as claimed in claim 1, is characterized in that, obtains the area to be processed that comprises shadow area based on preset face photo, is specifically: extracting the to-be-processed face image from the preset face photo based on the first preset cascade classifier; The to-be-processed region is extracted from the to-be-processed face image based on a second preset cascaded classifier. 3. The method according to claim 2, wherein the region to be processed is extracted from the face image to be processed based on a second preset cascade classifier, specifically: Obtaining a plurality of recognition results corresponding to the to-be-processed area from the to-be-processed face image based on the second preset cascade classifier; The to-be-processed area is extracted from a plurality of the identification results according to a preset verification rule, and the preset verification rule is determined according to a preset positional relationship of the facial features or a preset proportional relationship of the facial features. 4. The method of claim 1, wherein the shadow area is extracted from the to-be-processed area according to a preset pixel threshold, specifically: A binary image is segmented from the grayscale image of the region to be processed according to the preset pixel threshold, which is determined according to the Otsu method; The shaded area is extracted from the binary image. 5. The method according to claim 1, wherein gamma correction is performed on the shadow area based on a preset gamma coefficient, specifically: Iteratively adjusts the number of shaded pixels in the shaded area based on the preset gamma coefficient, until the number of shaded pixels is less than the preset number to determine a result of the gamma correction. 6. The method of claim 1, wherein generating a face photo with partial shadow removed according to the result area, specifically: generating a processed region based on replacing the shadow region with the result region; generating a processed face image based on replacing the processed region with the to-be-processed region; The partial shadow-removed face photo is generated based on replacing the processed face image with the to-be-processed face image. 7. A device for removing partial shadows of a face image, wherein the device comprises: a first obtaining module, configured to obtain a to-be-processed area including a shadow area based on a preset face photo, where the preset face photo includes a to-be-processed face image, and the to-be-processed face image includes the to-be-processed area; an extraction module, configured to extract the shadow area from the to-be-processed area according to a preset pixel threshold; a second obtaining module, configured to perform gamma correction on the shadow area based on a preset gamma coefficient, and obtain a result area based on the result of the gamma correction; A generating module is configured to generate a face photo with partial shadow removed according to the result area. 8. A computer-readable storage medium, characterized in that, an instruction is stored in the computer-readable storage medium, and when the instruction is executed on a terminal device, the terminal device is made to perform any one of claims 1-6. A method for removing partial shadows from a face image. 9. A computer program product, characterized in that, when the computer program product runs on a terminal device, the terminal device is made to execute the method for removing partial shadows from a face image according to any one of claims 1-6 .

Images