Image processing method and apparatus, and device

1. A method comprising:

capturing an image;

identifying a main object area or a background area in the image based on a category of an object in the image and using a neural network, wherein the background area is an area of the image other than the main object area, and wherein the identifying comprises:

performing, through the neural network, semantic segmentation on the image to obtain a plurality of masks, wherein the masks correspond to different object categories;

obtaining a target mask from the masks based on priorities of the object categories;

identifying, as the main object area, a first image area corresponding to the target mask; and

identifying, as the background area, a second image area corresponding to remaining masks of the masks;

when a terminal identifies a main subject and a background in the image, retaining a color of the main object area in the image; and performing black and white processing or blurring processing on the background area in the image;

when the terminal identifies only the background, performing the black and white processing or blurring processing on the background area; and

generating a target image or a target video based on the image.

2. The method of claim 1 , wherein a first mask of a first object category identifies the main subject area, wherein a second mask of a second object category identifies the background area, wherein the first object category comprises at least one of a person, an animal, a plant, a vehicle, or another preset object category, wherein the second object category is the background, and wherein the method further comprises determining the first mask and the second mask based on the neural network.

3. The method of claim 1 , further comprising obtaining the neural network based on a training data set for training, wherein the training data set comprises input images and mask graphs of a plurality of application scenarios.

4. The method of claim 2 , wherein the neural network identifies at least two object categories of the person, the animal, the plant, the vehicle, clothes, or the other preset object category.

5. The method of claim 2 , wherein a first quantity of pixels that are in the image and that correspond to the first mask is greater than a second quantity of pixels that are in the image and that correspond to a third mask of a third object category, and wherein the third object category is different from the second object category.

6. The method of claim 2 , wherein the first mask has a main object determining priority greater than a determining priority of another mask corresponding to another object category.

7. The method of claim 1 , wherein the main object area comprises a plurality of individuals, and wherein the individuals belong to a same object category or different object categories.

8. The method of claim 1 , further comprising obtaining the main object area by performing pixel-level segmentation on the object using the neural network.

9. An image processing apparatus, comprising:

a processor; and

a memory coupled to the processor and configured to store instructions that, when executed by the processor, cause the image processing apparatus to be configured to:

capture an image;

identify a main object area or a background area in the image based on a category of an object in the image using a neural network, wherein the background area is an area of the image other than the main object area, and wherein to identify the main object area or the background area, the instructions further cause the image processing apparatus to be configured to:

perform, through the neural network, semantic segmentation on the image to obtain a plurality of masks, wherein the masks correspond to different object categories;

determine a target mask from the masks based on priorities of the object categories;

identify, as the main object area, a first image area corresponding to the target mask; and

identify, as the background area, a second image area corresponding to remaining masks of the masks;

when the processor identifies a main subject and a background in the image,

retain a color of the main object area in the image; and

perform black and white processing or blurring processing on the background area in the image;

when the processor identifies only the background, perform the black and white processing or blurring processing on the background area; and

generate a target image or a target video based on the image after processing.

10. The image processing apparatus of claim 9 , wherein a first mask of a first object category identifies the main subject area, wherein a second mask of a second object category identifies the background area, wherein the first object category comprises at least one of a person, an animal, a plant, a vehicle, or another preset object category, wherein the second object category is the background, and wherein the instructions further cause the image processing apparatus to be configured to determine the first mask and the second mask based on the neural network.

11. The image processing apparatus of claim 9 , wherein the instructions further cause the image processing apparatus to be configured to obtain the neural network is based on a training data set for training, wherein the training data set comprises input images and mask graphs of a plurality of application scenarios.

12. The image processing apparatus of claim 10 , wherein the neural network identifies at least two object categories of the person, the animal, the plant, the vehicle, clothes, or the other preset object category.

13. The image processing apparatus of claim 10 , wherein a first quantity of pixels that are in the image and that correspond to the first mask is greater than a second quantity of pixels that are in the image and that correspond to a third mask of a third object category, and wherein the third object category is different from the second object category.

14. The image processing apparatus of claim 10 , wherein the first mask has a main object determining priority greater than a determining priority of another mask corresponding to another object category.

15. The image processing apparatus of claim 9 , wherein the main object area comprises a plurality of individuals, and wherein the individuals belong to a same object category or different object categories.

16. The image processing apparatus of claim 9 , wherein the instructions further cause the image processing apparatus to be configured to obtain the main object area by performing pixel-level segmentation on the object using the neural network.

17. A terminal device, wherein the terminal device comprises:

a camera configured to capture an image;

a processor coupled to the camera; and

a memory coupled to the processor and configured to store instructions that, when executed by the processor, cause the terminal device to be configured to:

capture an image;

identify a main object area or a background area in the image based on a category of an object in the image using a neural network, wherein the background area is an area of the image other than the main object area, and wherein to identify the main object area or the background area, the instructions that, when executed by the processor, cause the terminal device to be configured to:

perform, through the neural network, semantic segmentation on the image to obtain a plurality of masks, wherein the masks correspond to different object categories;

determine a target mask from the masks based on priorities of the object categories;

identify, as the main object area, a first image area corresponding to the target mask; and

identify, as the background area, a second image area corresponding to remaining masks of the masks;

when the terminal device identifies a main subject and a background in the image,

retain a color of the main object area in the image; and

perform black and white processing or blurring processing on the background area in the image;

when the terminal device identifies only the background, perform the black and white processing or blurring processing on the background area; and

generate a target image or a target video based on the image after processing.

18. The terminal device of claim 17 , wherein the terminal device further comprises an antenna system configured to receive and send a wireless communication signal under control of the processor to implement wireless communication with a mobile communications network that is one or more of a Global System for Mobile Communications (GSM) network, a code division multiple access (CDMA) network, a 3rd generation (3G) network, a 4th generation (4G) network, a 5th generation (5G) network, a frequency division multiple access (FDMA) network, a time division multiple access (TDMA) network, a primary domain controller (PDC) network, a total access communication system (TACS) network, an Advanced Mobile Phone System (AMPS) network, a wideband CDMA (WCDMA) network, a time domain synchronous CDMA (TDSCDMA) network, a WI-FI network, or a Long-Term Evolution (LTE) network.

19. An image processing apparatus, comprising:

a processor; and

a memory coupled to the processor and configured to store instructions that, when executed by the processor, cause the image processing apparatus to be configured to:

photograph a video;

identify a target area or a background area in a video image of the video based on a category of an object in the video image using a neural network, wherein the back round area is an area of the video image other than the target area, and wherein to identify the target area or the background area, the instructions that, when executed by the processor, cause the image processing apparatus to be configured to:

perform, through the neural network, semantic segmentation on the video image to obtain a plurality of masks, wherein the masks correspond to different object categories;

determine a target mask from the masks based on priorities of the object categories;

identify, as the target area, a first image area corresponding to the target mask; and

identify, as the background area, a second image area corresponding to remaining masks of the masks;

determine a main object in the video image of the video; and

retain a color of the target area in the video image, wherein the target area corresponds to the main object; and

perform black and white processing on the background area in the video image to obtain a target video,

wherein the background area is an area of the video image other than the target area.

20. The image processing apparatus of claim 19 , wherein the instructions further cause the image processing apparatus to be configured to perform filter processing on the background area, and wherein the filter processing comprises at least one of bokeh processing, darkening processing, retro processing, film processing, or texture superimposition processing.