Image annotation method, device, computer equipment and readable storage medium

Abstract

The invention relates to an image labeling method, device, computer equipment and readable storage medium. The method may include: marking the same image to be marked according to a pre-trained odd number of image labeling models, and obtaining corresponding odd number of first labeling images; the image labeling model is used to label the target object in the image; According to the odd number of first labeled images, a target labeled image of the image to be labeled is determined. In this image labeling method, since the odd number of first labeled images is obtained according to the pre-trained odd number of image labeling models, the labeling efficiency of the image to be labeled is improved. In addition, the computer device is determined according to the obtained odd number of first labeled images. The target labeling image of the image to be labeled avoids the problem of low quality labeling of the image to be labeling caused by the lack of personal experience of the labeler or the imperfection of the existing rules, and improves the labeling accuracy of the determined target labeling image.

Description

technical field [0001] The present invention relates to the image field, in particular to an image labeling method, device, computer equipment and readable storage medium. Background technique [0002] In recent years, with the further development of high-performance image processors (Graphics Processing Unit, GPU) and deep learning technology, more and more users use deep learning technology to detect objects in two-dimensional images. The first step of deep learning-based two-dimensional image object detection technology is to accurately label a large amount of two-dimensional image data to provide a large amount of sample data for deep learning. [0003] In the traditional image annotation method, the annotator judges the position of the object in the two-dimensional image based on personal experience and some existing rules, and uses the existing annotation tool to drag the mouse on the two-dimensional image and mark it with a bounding box. Outline the object and mark t...

AI Technical Summary

Problems solved by technology

[0005] Based on this, it is necessary to provide an image annotation method for the problems that the traditional image annotation method consumes a lot of manpower and time, the annotation efficiency is low, and the annotation quality may be low due to the lack of personal experience of the annotator or the imperfect existing rules. Method, device, computer device and readable storage medium

Images