Search

CN-122002138-A - Image processing method, apparatus, storage medium, and program product for large model

CN122002138ACN 122002138 ACN122002138 ACN 122002138ACN-122002138-A

Abstract

The embodiment of the disclosure provides an image processing method, equipment, a storage medium and a program product for a large model, which relate to the technical field of large data and image processing, and the method comprises the steps of acquiring a window image of a shooting window of a camera; cutting and removing an invalid region in the window image to obtain an effective window image, adjusting a first exposure parameter of the camera according to the effective window image, controlling the camera to shoot according to the first exposure parameter to obtain at least one image to be identified, inputting the at least one image to be identified into a large model, and participating in large model processing. By cutting the window image, adjusting exposure parameters according to the cut effective window image, and acquiring an image input into the large model by adopting the exposure parameters, the problem that the prediction accuracy of the large model is reduced under the condition of overexposure or underexposure of the image is solved.

Inventors

  • TAO XUJIE
  • LIU SHUO
  • GUO JIA
  • LIU MENGQIAN

Assignees

  • 北京字跳网络技术有限公司

Dates

Publication Date
20260508
Application Date
20241104

Claims (16)

  1. 1. An image processing method for a large model, applied to a terminal device having a camera, comprising: Acquiring a window image of a shooting window of a camera; cutting and removing an invalid region in the window image to obtain an effective window image; according to the effective window image, adjusting a first exposure parameter of the camera; Controlling a camera to shoot according to the first exposure parameters to obtain at least one image to be identified; and inputting the at least one image to be identified into a large model, and participating in large model processing.
  2. 2. The method according to claim 1, wherein the terminal device has at least one camera in an operational state; correspondingly, the cropping removes the invalid region in the window image to obtain an effective window image, which comprises the following steps: Determining an invalid region in the window image according to the shooting position of the camera in the at least one running state; And removing the invalid region in the window image to obtain an effective window image.
  3. 3. The method according to claim 2, characterized in that the terminal device has two cameras in operation; Correspondingly, the determining the invalid area in the window image according to the shooting position of the camera in the at least one running state comprises the following steps: And determining an invalid region in the window image according to the shooting positions of the cameras in the two running states.
  4. 4. The method of claim 1, wherein the cropping removes invalid regions in the window image to obtain a valid window image, comprising: inputting the window image into an interest point determination model to obtain an interest point region output by the interest point determination model; and cutting the region outside the interest point region in the window image to obtain an effective window image.
  5. 5. The method of claim 4, wherein the point of interest determination model is trained by: acquiring marked image data and language description corresponding to the marked image data; And inputting the marked image data and the corresponding language description into a large model to be trained so as to enable the large model to be trained to conduct comparison learning, and obtaining the interest point determination model.
  6. 6. The method of claim 1, wherein the cropping removes invalid regions in the window image to obtain a valid window image, comprising: inputting the window image and the preset self equipment type into an interest point determination model to obtain an interest point region output by the interest point determination model; and cutting the region outside the interest point region in the window image to obtain an effective window image.
  7. 7. The method of claim 1, wherein adjusting the first exposure parameter of the camera based on the active window image comprises: And inputting the effective window image into an exposure parameter determination model to obtain a first exposure parameter output by the exposure parameter determination model.
  8. 8. The method of claim 1, further comprising, after the capturing of the view image of the camera's captured view: receiving semantic data input by a user; determining a second exposure parameter according to the window image and the semantic data; controlling a camera to shoot according to the second exposure parameters to obtain at least one image to be identified; and inputting the at least one image to be identified and the semantic data into a large model to participate in large model processing.
  9. 9. The method according to any one of claims 1 to 8, further comprising, after said controlling the camera to take the image according to the first exposure parameter, at least one image to be identified: cutting and removing an invalid region in the image to be identified to obtain an effective image to be identified; Performing histogram equalization on the effective image to be identified to obtain an equalized image to be identified; And inputting the balanced image to be identified into a large model, and participating in large model processing.
  10. 10. The method of any one of claims 1 to 8, further comprising, after said adjusting the first exposure parameter of the camera based on the active window image: Determining whether the exposure is overexposed or underexposed; if overexposure or underexposure is performed, outputting prompt information for adjusting shooting position and optimizing exposure.
  11. 11. The method according to any one of claims 1 to 8, further comprising, after the capturing of the view image of the captured view of the camera: Acquiring a hardware resource state; correspondingly, the controlling the camera to shoot according to the first exposure parameter to obtain at least one image to be identified includes: If the hardware resource state does not meet the preset requirement, controlling a camera to shoot according to the first exposure parameter to obtain an image to be identified; And if the hardware resource state meets the preset requirement, controlling the camera to shoot according to the first exposure parameter to obtain at least two images to be identified.
  12. 12. The method of claim 11, wherein the hardware resource status comprises at least one of power, bandwidth, and computing resources; Correspondingly, if the hardware resource state does not meet the preset requirement, controlling the camera to shoot according to the first exposure parameter to obtain an image to be identified, including: if the electric quantity is smaller than the preset electric quantity threshold value, controlling the camera to shoot according to the first exposure parameter to obtain an image to be identified, or, If the bandwidth is smaller than the preset bandwidth threshold, the camera is controlled to shoot according to the first exposure parameter to obtain an image to be identified, or, And if the computing resource is smaller than a preset computing resource threshold, controlling the camera to shoot according to the first exposure parameter to obtain an image to be identified.
  13. 13. An image processing apparatus for a large model, comprising: the image acquisition module is used for acquiring a window image of a shooting window of the camera; the image cutting module is used for cutting and removing an invalid area in the window image to obtain an effective window image; the parameter adjusting module is used for adjusting a first exposure parameter of the camera according to the effective window image; the image shooting module is used for controlling the camera to shoot according to the first exposure parameters to obtain at least one image to be identified; and the model processing module is used for inputting the at least one image to be identified into the large model and participating in the large model processing.
  14. 14. The electronic equipment is characterized by comprising a camera, a processor and a memory; The memory stores computer-executable instructions; The processor controls the camera, executes computer-executable instructions stored in the memory, and causes the processor to perform the image processing method for a large model as claimed in any one of claims 1 to 12.
  15. 15. A computer-readable storage medium, in which computer-executable instructions are stored which, when executed by a processor, implement the image processing method for large models according to any one of claims 1 to 12.
  16. 16. A computer program product comprising a computer program which, when executed by a processor, implements a method for image processing of large models according to any of claims 1 to 12.

Description

Image processing method, apparatus, storage medium, and program product for large model Technical Field The embodiment of the disclosure relates to the technical field of big data and image processing, in particular to an image processing method, device, storage medium and program product for a big model. Background With the development of artificial intelligence, large models have been provided with the ability to combine picture generation results. But this capability is affected by the quality of the input image. The high-quality image not only can remarkably improve the prediction, analysis and prediction capabilities of the model, but also can ensure the accuracy of the result. Currently, in some scenes (for example, in a wearable terminal device), a user cannot check the shooting condition of an image in time, and cannot determine whether the image is overexposed or underexposed. However, the inventors found that the related art has at least a technical problem in that the prediction accuracy of a large model is lowered in the case of overexposure or underexposure of an image. Disclosure of Invention Embodiments of the present disclosure provide an image processing method, apparatus, storage medium, and program product for a large model to solve the problem that the prediction accuracy of the large model is degraded in the case of overexposure or underexposure of an image. In a first aspect, an embodiment of the present disclosure provides an image processing method for a large model, including obtaining a window image of a shooting window of a camera, cutting an invalid region in the window image to obtain an effective window image, adjusting a first exposure parameter of the camera according to the effective window image, controlling the camera to shoot according to the first exposure parameter to obtain at least one image to be identified, and inputting the at least one image to be identified into the large model to participate in large model processing. In a second aspect, an embodiment of the disclosure provides an image processing device for a large model, which includes an image acquisition module for acquiring a window image of a shooting window of a camera, an image cutting module for cutting an invalid region in the window image to obtain an effective window image, a parameter adjustment module for adjusting a first exposure parameter of the camera according to the effective window image, an image shooting module for controlling the camera to shoot according to the first exposure parameter to obtain at least one image to be identified, and a model processing module for inputting the at least one image to be identified into the large model to participate in large model processing. In a third aspect, embodiments of the present disclosure provide an electronic device comprising a processor and a memory, the memory storing computer-executable instructions, the processor executing the computer-executable instructions stored in the memory, such that at least one processor performs the image processing method for a large model as described above in the first aspect and the various possible designs of the first aspect. In a fourth aspect, embodiments of the present disclosure provide a computer-readable storage medium having stored therein computer-executable instructions which, when executed by a processor, implement the image processing method for large models as described above in the first aspect and the various possible designs of the first aspect. In a fifth aspect, embodiments of the present disclosure provide a computer program product comprising a computer program which, when executed by a processor, implements the image processing method for large models as described in the first aspect and the various possible designs of the first aspect. According to the image processing method, the device, the storage medium and the program product for the large model, the method is used for obtaining the effective window image by acquiring the window image, cutting the ineffective area in the window image, determining the exposure parameters by adopting the effective window image, shooting the image to be identified by adopting the determined exposure parameters, inputting the image to be identified into the large model for subsequent processing, avoiding the influence of the ineffective area on exposure, and increasing the exposure accuracy degree under the conditions that a user does not need to observe the exposure condition and manually adjust the exposure parameters, so that the accuracy degree of the content output by the large model is higher. Drawings In order to more clearly illustrate the embodiments of the present disclosure or the solutions in the prior art, a brief description will be given below of the drawings that are needed in the embodiments or the description of the prior art, it being obvious that the drawings in the following description are some embodiments of the present disclosu