Feedback labeling method and device for machine learning object detection result and storage medium

Abstract

The invention discloses a feedback labeling method and device for a machine learning object detection result. The feedback labeling method and device are used for solving the problem that feedback onthe machine learning object detection result is lacked in the related technology. The labeling device displays current frame result image data of machine learning object detection, the image data comprises at least one first labeling box and at least one second labeling box. The first labeling box is a manually labeled labeling box, and the second labeling box is a labeling box for executing object detection labeling through machine learning. An input annotation box comparison parameter is received, wherein the annotation box comparison parameters comprise whether a first annotation box and asecond annotation box which have the same identifier coincide or not and the coincidence proportion of the first annotation box and the second annotation box; The result image data of the current frame and the received annotation box comparison parameters are stored in an associated manner.

Description

technical field [0001] The present application relates to the field of data labeling, in particular to a feedback labeling method, device, and storage medium for machine learning object detection results. Background technique [0002] In related technologies, the manually labeled image data is usually used as training data to provide to the machine learning module. The machine learning module performs object detection and learning on the image data with the set parameters. After the training of the machine learning module is determined to be completed, the executable A machine learning module for object detection. The trained machine learning module performs object detection on the input image data and outputs the resulting image data. [0003] Whether the object detection result output by the machine learning module is accurate is usually identified and judged by the developer. The training of the machine learning module can be done when the developer deems the object det...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present application provides a method and device for feedback labeling of machine learning object detection results to solve the problem of lack of feedback on machine learning object detection results in related technologies

Images