Segmentation method of eye region based on fuzzy logic system for classifying open and closed eyes

Abstract

The classification of eye openness and closure has been researched in various fields, e.g., driver drowsiness detection, physiological status analysis, and eye fatigue measurement. For a classification with high accuracy, accurate segmentation of the eye region is required. Most previous research used the segmentation method by image binarization on the basis that the eyeball is darker than skin, but the performance of this approach is frequently affected by thick eyelashes or shadows around the eye. Thus, we propose a fuzzy-based method for classifying eye openness and closure. First, the proposed method uses I and K color information from the HSI and CMYK color spaces, respectively, for eye segmentation. Second, the eye region is binarized using the fuzzy logic system based on I and K inputs, which is less affected by eyelashes and shadows around the eye. The combined image of I and K pixels is obtained through the fuzzy logic system. Third, in order to reflect the effect by all the inference values on calculating the output score of the fuzzy system, we use the revised weighted average method, where all the rectangular regions by all the inference values are considered for calculating the output score. Fourth, the classification of eye openness or closure is successfully made by the proposed fuzzy-based method with eye images of low resolution which are captured in the environment of people watching TV at a distance. By using the fuzzy logic system, our method does not require the additional procedure of training irrespective of the chosen database. Experimental results with two databases of eye images show that our method is superior to previous approaches.

A Study of Deep CNN-Based Classification of Open and Closed Eyes Using a Visible Light Camera Sensor

“…For the next experiment, we compared the testing accuracies for classifying open and closed eye images with the proposed deep residual CNN to other methods. For comparison, we used HOG-SVM [ 31 ], fuzzy system-based method [ 17 ], the conventional CNN models of AlexNet [ 34 ], VGG face [ 36 ], and GoogLeNet [ 38 ]. The same databases were used for our method and the other methods, and the average accuracies were measured using two-fold cross validation.…”

“…For the training of AlexNet, VGG face, and GoogLeNet, the mini-batch size was set to 25, and the testing accuracies were measured for 10, 20, and 30 training epochs. For non-CNN-based methods, we used HOG-SVM method [ 31 ] and fuzzy system-based method [ 17 ] for comparison. In case of using HOG-SVM method, from the input images, the features were extracted by HOG, and the classification of open and closed eye images was performed based on SVM using a radial basis function (RBF).…”

“…In case of using HOG-SVM method, from the input images, the features were extracted by HOG, and the classification of open and closed eye images was performed based on SVM using a radial basis function (RBF). The fuzzy system-based method [ 17 ] uses the standard deviation of vertical pixel length calculated from binarized image for classification. The binarized image is obtained by fuzzy-system based segmentation.…”

“…One of the most famous examples of a single-image-based (non-training based) method is iris detection. There have been studies on methods using the difference between the maximum iris location in a vertical histogram, assuming the existence of irises in an image, and the vertical location of the iris centers [ 14 ], a method to determine when the pupils are present via sub-block-based template matching and ellipse-fitting [ 15 ], template-matching-based methods [ 16 ], and a fuzzy-system-based method using eye segmentation [ 17 ]. The main advantage of these methods is that it is possible to distinguish between open and closed eyes in a single image without an additional training process.…”

Deep learning and blockchain fusion for detecting driver's behavior in smart vehicles

“…In computer vision, distracted behavior of the driver has been detected using two different approaches. First one is based on the traditional hand crafted features and, then training these extracted features on machine 4 . These approaches mostly use the features of driver eyes, head and face for predicting the anonymous behavior of the driver.…”