All-weather eye circumference identification method based on cascade super-resolution

Abstract

The invention belongs to the technical field of mode recognition and digital image processing, and particularly relates to an all-weather eye circumference recognition method based on cascade super-resolution. The method comprises the following steps of using a multispectral camera for collecting eye circumference images; pre-processing a limited eye circumference data set and performing sample expansion. carrying out image super-resolution operation on the periocular data set by adopting a convolutional neural network based on deep learning so as to amplify the periocular image; carrying outimage restoration on the processed periocular data set by using a deep learning-based image deconvolution technology; based on the eye circumference data obtained through image restoration, adopting Laplace image sharpening enhancement; based on the obtained periocular data set, constructing a novel stacked neural network model by adopting a deep learning theory; dividing the periocular data set into a training set and a test set; and for the divided test set, calculating features through the trained convolutional neural network model. The method provided by the invention has relatively high robustness and good generalization performance.

Description

technical field [0001] The invention belongs to the technical fields of pattern recognition and digital image processing, and in particular relates to an all-weather eye circumference recognition method based on cascade super-resolution. Background technique [0002] Peripheral recognition is a new biometric mode with specific advantages that has emerged in recent years. For example, it can be used as a supplement to face recognition, and has been proven to have the most informative facial parts when the face is occluded. In addition, eye recognition can also be considered as an alternative when iris recognition fails, because iris recognition requires the acquisition of very high-quality iris images, which are usually not met at long distances. [0003] However, so far, researchers have only focused on visible light-based eye recognition technology, which usually does not perform well in harsh climates and environments, such as unconstrained situations and harsh climates s...

AI Technical Summary

Problems solved by technology

These traditional algorithms are all manually designed. Generally, the design is cumbersome and complicated, and they have good performance for specific situations but poor robustness. They are not good at solving problems such as lighting changes.

Images