Training image expansion method, training method and related device

Abstract

The invention provides a training image expansion method, a training method and a related device, and relates to the field of pedestrian recognition of machine learning. The training image expansion method comprises the steps of obtaining a plurality of to-be-converted images of a pedestrian; wherein the plurality of to-be-converted images comprise at least two pedestrian color data of pedestrians; obtaining a color data generation model according to the plurality of to-be-converted images; wherein the color data generation model is a model obtained by training at least two kinds of pedestriancolor data under a generative adversarial network; inputting the plurality of images to be converted into a color data generation model to obtain a plurality of target color images; and taking the to-be-converted image and the plurality of target color images as a color training set. By using the data expansion method provided by the invention, more specific color data can be obtained on the basis of real color data, so that the training data requirement of color recognition model training is met, and the accuracy of color recognition is improved.

Description

technical field [0001] The present application relates to the field of pedestrian recognition of machine learning, and specifically relates to an expansion method of training images, a training method and related devices. Background technique [0002] The current method of pedestrian clothing color recognition based on deep learning relies on a large amount of labeled data, which requires a lot of manpower on the data. [0003] However, for some niche training data, it is not possible to collect enough training data, such as the pedestrian data of the brown coat, which hardly appears in a surveillance video. Due to the lack of these training data, the color recognition model is not good at these niche colors. Recognition is always suboptimal. Therefore, how to quickly expand the color annotation data is a problem that needs to be solved at present. Contents of the invention [0004] In order to overcome the above-mentioned shortcomings in the prior art, one object of the...

AI Technical Summary

Problems solved by technology

[0002] The current method of pedestrian clothing color recognition based on deep learning relies on a large amount of labeled data and requires a lot of manpower on the data.

[0003] However, for some niche training data, it is not possible to collect enough training data, such as the pedestrian data of the brown coat, which hardly appears in a surveillance video. Due to the lack of these training data, the color recognition model is not good at these niche colors. Recognition is not always ideal

Images