An image capturing method, device, equipment and storage medium
An image capture and image technology, which is applied in the field of computer software applications, can solve problems such as the complexity of portrait shooting, and achieve the effect of improving shooting effects and user experience
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0033] see figure 1 , the present embodiment provides an image capturing method, the method comprising the following steps:
[0034] S110. Obtain a bounding box of a lens tracking target in an image to be captured.
[0035] When shooting images, in order to achieve a better composition effect, the target to be photographed or the lens tracking target is usually placed in the center of the image as much as possible. Therefore, before adjusting the lens movement, it is necessary to determine the position of the tracking target in the image. Here The lens tracking target refers to the main shooting target that needs to be kept in the lens at all times, such as people, pets and other photographic materials. In this embodiment, a bounding box is used to determine the position of the tracking target. The bounding box refers to the area corresponding to the frame where the tracking target appears in the image to be captured, and generally has a rectangular outer frame shape that is ...
Embodiment 2
[0053] image 3 It is a schematic flowchart of an image capturing method provided in Embodiment 2 of the present invention. This embodiment is implemented on the basis of Embodiment 1, as shown in image 3 As shown, before step S110 also includes:
[0054] Step S100, obtain a pre-trained reference model based on deep convolutional neural network training.
[0055] In some embodiments, such as Figure 4 As shown, step S100, based on deep convolutional neural network training to obtain a pre-trained reference model (that is, the specific training process of the reference model) includes steps S310-step S360, specifically as follows:
[0056] S310. Obtain training images and corresponding labeled data from a preset image data set, where the labeled data includes bounding box information and key point information of the target.
[0057] In this embodiment, multiple training images are pre-set in the image data set, and the type of training images can be selected according to di...
Embodiment 3
[0128] Such as Image 6 As shown, this embodiment provides an image capture device 500, including:
[0129] A bounding box acquisition module 510, configured to acquire the bounding box of the lens tracking target in the image to be captured;
[0130] A reference position prediction module 520, configured to use a pre-trained reference model to predict the first reference position of the image to be captured;
[0131] A lens offset determining module 530, configured to determine a lens shift offset according to each pixel position in the bounding box and the first reference position.
[0132] In this embodiment, the bounding box acquisition module 510 may acquire multiple corresponding bounding boxes according to the number of tracking targets in the image to be captured.
[0133] In this example, if Figure 7 As shown, the reference position prediction module 520 also includes a model training submodule 521, which is used to obtain a trained reference model based on deep c...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More