Object Segmentation Method, Device, and Computing Equipment Based on Multi-level Local Area Fusion

Abstract

The invention discloses an object segmentation method, device and computing device based on multi-level local region fusion, belonging to the technical field of computer vision, wherein the method includes: for an image to be processed, selecting multiple local candidate regions; image segmentation processing is performed on each local candidate region, and the binary segmentation mask of the local candidate region is predicted; image classification processing is performed on each local candidate region, and the object category to which the local candidate region belongs is predicted; according to each The object category to which the local candidate region belongs and the binary segmentation mask of each local candidate region are merged to obtain the object segmentation image. The invention can segment the individual objects and determine their precise boundaries while detecting the objects. After obtaining the segmentation result of the local candidate area, the present invention uses an effective local area fusion method to obtain a better optimization result.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to an object segmentation method, device and computing device based on multi-level local region fusion. Background technique [0002] Image segmentation is a basic problem in the field of image processing, and has a wide range of applications in object recognition, robot navigation, scene understanding and other fields. Among them, object segmentation is a more essential problem than object detection and scene segmentation. Object detection expects to obtain the approximate position of each individual object. Usually, object detection marks the detected object through a rectangular detection frame, but cannot determine the boundary of the object; the scene is segmented into each type of scene to predict the category and the precise boundary, which can be used for different types of objects Make predictions and determine boundaries, but cannot distinguish between individuals...

AI Technical Summary

Problems solved by technology

Object detection expects to obtain the approximate position of each individual object. Usually, object detection marks the detected object through a rectangular detection frame, but cannot determine the boundary of the object; the scene is segmented into each type of scene to predict the category and the precise boundary, which can be used for different types of objects Make predictions and determine boundaries, but cannot distinguish between different individuals of the same class

[0003] The method of object segmentation in the prior art is mainly to obtain the candidates of the detection frame based on object detection, and then obtain the boundary through the segmentation method. The segmentation effect of this method is poor and the accuracy rate is not high.

Images