A method and system for interactive image text annotation

Abstract

The invention provides an interactive image text labeling method, which comprises the following steps: a server end receives an image and / or text data sample labeled by a user, and trains a model at the server end based on the sample. The trained model pre-annotates data which is newly loaded to the server end. The pre-annotated data is sent to the client for correction; And the server receives the corrected data to iteratively update the training model. The invention invokes the depth learning algorithm, can carry out on-line or off-line training on the labeled data samples to update the labeling model, continuously optimizes the labeling process, greatly improves the labeling efficiency, and saves human resources and time cost for customers.

Description

technical field [0001] The present invention generally relates to the field of computers, and more specifically, relates to an interactive image text labeling method and system. Background technique [0002] The development of artificial intelligence technology has promoted the technological change and model innovation of society, and has given birth to the development of emerging technologies including robots, autonomous driving, voice recognition, online consultation and intelligent medical care. At present, the biggest driving force for the development of artificial intelligence technology is machine learning or deep learning technology. The current mainstream is deep learning technology, which is widely used in the fields of computer vision and natural language processing. The basis is that it is necessary to add annotation information to data files, including image target detection. , Target tracking, image classification, image segmentation and other image annotation t...

AI Technical Summary

Problems solved by technology

Text annotation tools often need to use a lot of manpower and material resources to annotate the collected corpus. After the annotation is completed, the model file is trained. If new vocabulary appears, the annotated corpus often needs to be re-annotated. Such multiple reworks seriously reduce the user's work efficiency.

[0004] To sum up, although there are many annotation tools in the field of computer vision and natural language processing, these annotation tools mainly use manual annotation methods, which is time-consuming, laborious, and inefficient; different labeling tasks, which brings inconvenience to personnel who take care of multiple labeling tasks

Images