Method and device for target boundary segmentation and background noise suppression based on neural network

Abstract

The present application provides a neural network-based target boundary segmentation and background noise suppression method and device, the method comprising: acquiring an image and determining the target to be detected and the background target; respectively performing bounding frame segmentation on the target to be detected and the background target to obtain the target to be detected Detect the bounding box of the target and the background target; perform edge detection of the target to be detected; output the target to be detected. A computer-readable medium stores computer-readable instructions thereon, and when the computer-readable instructions can be executed by a processor, the processor can implement the above-mentioned method. The device includes: a processor; a computer-readable medium for storing computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor implements the above method. When the background target and the target to be detected are closely connected with each other, even if the bounding box contains a lot of background noise, the application can still clearly calibrate the complete and accurate boundary of the target to be detected.

Description

technical field [0001] The present application relates to the field of computers, in particular to methods and devices for boundary detection and background noise suppression based on neural networks. Background technique [0002] When using a neural network-based object boundary segmentation method for such as figure 1 The multiple targets of the dense distribution shown are divided by circumscribed rectangles, and the following is obtained: figure 2 The results shown. However, this method has the following problem: the segmentation rectangle contains many other background objects besides the object. by image 3 For example, A is the target to be detected, B and C are background targets or interference targets. Such segmentation results cannot meet the precise positioning requirements of the target object, and the precise positioning results are as follows: Figure 4 As shown, the rectangular frame surrounding the target A to be detected is the precise positioning fra...

AI Technical Summary

Problems solved by technology

[0004] One purpose of this application is to provide a method and device based on neural network boundary detection and background noise suppression to solve the problem that the circumscribed rectangular frame contains a large amount of background noise in the prior art. When the interfering target and the target to be detected are closely connected to each other, then The technical problem that the complete and accurate boundary of the target to be detected cannot be clearly calibrated

Images