Unlock instant, AI-driven research and patent intelligence for your innovation.

Target area determination model training method, device and computer-readable storage medium

A target area and model training technology, applied in computer parts, computing, biological neural network models, etc., can solve problems such as misjudgment of target areas, and achieve the effect of improving recognition accuracy

Active Publication Date: 2021-06-22
BEIJING BYTEDANCE NETWORK TECH CO LTD
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem solved by the present disclosure is to provide a target area determination model training method to at least partly solve the technical problem that the target area is misjudged in the prior art

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Target area determination model training method, device and computer-readable storage medium
  • Target area determination model training method, device and computer-readable storage medium
  • Target area determination model training method, device and computer-readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0085] In order to solve the technical problem of a low correct recognition rate of a target area in the prior art, an embodiment of the present disclosure provides a method for training a target area determination model. Such as Figure 1a As shown, the target area determination model training method mainly includes the following steps S11 to S14. in:

[0086] Step S11: Perform target area recognition on the sample image to obtain an image area including the target area.

[0087] Wherein, the target area may be a license plate area.

[0088] Wherein, the sample image is an image input into the convolutional neural network as a training sample for training. Specifically, it may be obtained through a camera of a terminal device, or stored in a local database in advance, and obtained from a local database.

[0089] Specifically, an existing convolutional neural network model may be used to identify the sample image to obtain an image area including the target area.

[0090] ...

Embodiment 2

[0114] In order to solve the technical problem of low accuracy rate of target area determination in the prior art, the embodiment of the present disclosure also provides a target area determination method, such as figure 2 shown, including:

[0115] S21: Perform target area identification on the currently input video frame to obtain an image area including the target area.

[0116] Wherein, the input video frame can be obtained in real time through the camera, or a pre-stored video image can be obtained locally.

[0117] Wherein, the target area may be a license plate area, and the corresponding target area is a rectangular area.

[0118] Specifically, an existing convolutional neural network model may be used to perform preliminary identification on the input video frame to obtain an image area including the target area.

[0119] S22: Scale the image area to a fixed size.

[0120] Wherein, the fixed size is consistent with the image area of ​​the training sample input dur...

Embodiment 3

[0141] In order to solve the technical problem of low target area determination accuracy in the prior art, an embodiment of the present disclosure provides a target area determination model training device. The device can execute the steps in the embodiment of the target area determination model training method described in the first embodiment above. Such as image 3 As shown, the device mainly includes: a sample identification module 31, a training set determination module 32, a training set input module 33 and a model training module 34; wherein,

[0142] The sample identification module 31 is used to identify the target area of ​​the sample image to obtain an image area including the target area;

[0143] The training set determination module 32 is used to scale the image area to a fixed size, a training sample set is composed of multiple image areas of the fixed size, and the image areas in the training sample set are marked with multiple key points, Wherein, the plural...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The disclosure discloses a target area determination model training method, device, electronic equipment and computer-readable storage medium. The method includes: identifying the target area on the sample image to obtain an image area containing the target area; a training sample set is composed of a plurality of fixed-sized image areas, and the image areas in the training sample set are marked with multiple key points. A key point is located in the target area; each training channel of the convolutional neural network is trained independently until the respective convergence conditions are met, and a target area determination model including multiple training channels is obtained; multiple training channels are used to predict multiple key points respectively The displacement relative to the reference point. The embodiments of the present disclosure respectively train the training sample set through multiple parallel training channels to obtain multiple training channels, and the multiple training channels are respectively used to predict the displacement of multiple key points relative to the reference point, thereby obtaining The target area can improve the accuracy of target area recognition.

Description

technical field [0001] The present disclosure relates to the technical field of target area determination model training, and in particular to a target area determination model training method, device and computer-readable storage medium. Background technique [0002] Many of the captured video images contain cars, and the images containing cars generally include license plates. Since the license plate involves privacy, it is necessary to process the license plate in the video image or use other images to cover the license plate. When processing the image containing the license plate, it is the key to recognize the license plate area in the image. [0003] For the recognition of the license plate area in the prior art, a model is generally used to judge whether each pixel in the license plate is foreground or background, wherein the foreground is the license plate area to be recognized, and the background refers to the area in the image except the license plate. But using t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/32G06K9/46G06K9/62G06N3/04
CPCG06V10/25G06V10/462G06N3/045G06F18/2155
Inventor 朱延东王长虎
Owner BEIJING BYTEDANCE NETWORK TECH CO LTD