English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Human eye location plays a crucial part in machine vision fatigue detection. To address the difficulty of distinguishing eye and brow in traditional way of machine vision human eye location, this paper proposes an accurate human eye location way, which is based on geometrical characteristics of eye and brow, capable to avoid mistakes of eye and brow judgment, and makes locations fast and precisely. Firstly, with the most suitable threshold segmentation algorithm, separate the suspected human eye region from face through binary image; secondly, process the binary image of human face with mathematic morphology; lastly, make accurate human eye location by distinguishing eye and brow geometrical characteristics. As the experiments show, the method proposed in this paper can avoid the interference from brow and locate human eye accurately. 1 INSTRUCTION Statistical reports at home and abroad show the accidents caused by driver’s fatigue driving take up 14.9% of human injury accidents and 20.6% of fatal accidents. According to statistics, when traffic accident happens, if driver can react a half second earlier, 60% of those accidents can be avoided. Obviously, there is an urgent necessity to do fatigue detecting for drivers. For drivers, the no-contact machine vision fatigue detecting is the best preference, which judges driver’s fatigue degree in the light of facial information. As to the method, eye positioning is the most important part, thus it makes human eye positioning in machine vision fatigue of great importance. Generally, there are two considerations in the process of human eye detecting: one is real-time performance, the other is accuracy. The current human eye detecting methods are comparatively hardly to meet the two requirements at the same time. The method based on prior knowledge boasts good real-time, but it’s hard to master the degree of rules while making conclusion of experiences, which makes it hardly to be popularized. The template matching method enjoys higher accuracy compared with the former one, but it’s susceptible to posture, brow and external interference. The method based on statistics has high accuracy, but the demand of large quantity of samples makes it of complicated process and poor real-time performance. Gray projection method provides simple algorithm and good real-time, but it’s susceptible to environment and with weak robustness. After analyzing above methods and analyzing actual situation, this paper proposes a human eyelocation method which combines threshold segmentation algorithm and eye and brow characteristics. Firstly, process the extracted facial images with linear graying; secondly, separate suspected human eye region from background with threshold segmentation; lastly, make accurate eye location according to eye and brow characteristics after morphological processing. The algorithm flow chart goes as Fig.1. International Conference on Applied Science and Engineering Innovation (ASEI 2015) © 2015. The authors Published by Atlantis Press 1012 Fig.1 lgorithm flow chart 2 THRESHOLD SEGMENTATION EYE DETECTION METHOD 2.1 Facial linear graying processing The gray of human eye changes larger than other parts of face, which makes it possible to adopt the most suitable threshold method to detect human eye in human face region. Process the marked facial region with linear graying to get human face grayscale image. The process goes as equation (1): B G R B G R Gray 114 . 0 587 . 0 299 . 0       (1) R, G, and B represent three channel component values in RGB space. Out of increasing the contrast ratio of facial image, facial region is suggested to go through linear processing. The process goes as equation (2):             ) 255 ( , 255 ) 255 ( , Gray k if Gray k if Gray k Gray (2) K equals 2.8. The contrast ratio of brow, eye, and some other facial parts increases greatly after the facial image being processed with linear graying. 2.2 The most suitable threshold segmentation Pupil is the part of smallest facial gray, almost being black. But the gray of sclera, or to say eye white, is very large, almost being white. The main idea of this algorithm is to take advantage of eye gray changes, use iterative method to choose a most suitable threshold value, and then separate thesuspected human eye region from background, and finally achieve the goal to detect human eye. Firstly, preset an initial threshold 0 T . In this paper, the initial threshold is set to be the average value of max Gray and min Gray of facial region. 2 min max 0 Gray Gray T   (3) Take the image as two parts, detecting target and background. Then adopt iterative method to determine the average value of target threshold and background threshold, O Gray and B Gray . The equations go as follows:

Accurate Human eye Positioning Based on the Geometrical Characteristics of eye and Brow