Eye Status Based on Eyelid Detection: A Driver Assistance System

Abstract

Abstract. Fatigue and driver drowsiness monitoring is an important subject for designing driver assistance systems. The measurement of eye closure is a fundamental step for driver awareness detection. We propose a method which is based on eyelid detection and the measurement of the distance between the eyelids. First, the face and the eyes of the driver are localized. After extracting the eye region, the proposed algorithm detects eyelids and computes the percentage of eye closure. Experimental results are performed on the BioID database. Our comparisons show that the proposed method outperforms state-of-the-art methods.

Eyelid and Pupil Landmark Detection and Blink Estimation Based on Deformable Shape Models for Near-Field Infrared Video

“…The majority of studies on eyelid and blink detection are for frontal face images taken under normal light conditions and in the far field. They usually begin with face recognition and then estimate eye regions, cropped according to topography rules (Moriyama et al, 2006;Bacivarov et al, 2008;Yang et al, 2012;Mohanakrishnan et al, 2013;Daniluk et al, 2014;Yahyavi et al, 2016) or utilizing facial landmarks (Fridman et al, 2018). The eye images thus are in low resolution with clear views of only the iris and the eyelid contour.…”

“…The general principal of eyelid detection has been to utilize the difference of intensity or edges between the skin, iris, and sclera. Methods to date include finding the points which have maximum response to an upper eyelid and lower eyelid filter (Tan and Zhang, 2006;Daniluk et al, 2014); heuristically examining the polynomials fitted to eye corner points with each, pairs and triples of edge segment above the iris center for topography criteria (Sirohey et al, 2002); and searching among parabolas which pass through the eye corner points to maximize an objective function involving edges, intensity and area (Kuo and Hannah, FIGURE 1 | Why near-field IR eye activity estimation is difficult: unlike far-field non-IR eye images in high resolution [first row of (A)] (Alabort-i- Medina et al, 2014) and low resolution [second and third rows of (A)] (Mohanakrishnan et al, 2013), where the eyelash is barely visible and the intensity between skin, the sclera and iris is often distinctive, IR near-field eye images, (B) enhance the contrast between the pupil and its surroundings but weaken the contrast between skin, the sclera and iris. The close-up view highlights greater eye appearance difference due to ethnicity, blinking, motion blur, eyelash density, illumination changes, and camera positions than the far-field view.…”

“…Eye activity has been of great interest since observations of human attention and intention began (Duchowski, 2007). The types and applications of eye activity that have been investigated up to now include gaze direction as a pointing device for paralyzed people (Duchowski, 2007); pupil size, blink, and eye movement (fixation and saccade) for cognitive load measurement (Chen et al, 2011), emotion recognition (Lu et al, 2015), visual behavior change (Chen et al, 2013), human activity recognition (Bulling et al, 2011), and mental illness diagnosis (Vidal et al, 2012); eyelid closure for emotion recognition (Orozco et al, 2009), and fatigue detection (Yang et al, 2012;Daniluk et al, 2014). As opposed to these dynamic changes in eye components (eyelid opening, pupil size and location, blink length and depth) which form eye activities, static eye images are also of interest especially in biometrics.…”

Sistema Automático Para la Detección de Distracción y Somnolencia en Conductores por Medio de Características Visuales Robustas

“…Para entrenar el algoritmo se recolectaron imágenes positivas (el conductor está presente) e imágenes negativas (el conductor no está presente y, por tanto, se veúnicamente el interior del vehículo). Para la recolección de imágenes positivas, se tomaron 1000 imágenes de las siguientes bases de datos: a) YawDD (Abtahi et al (2014)), b) RS-DMV (Nuevo et al (2010)), c) Set 11 (Daniluk et al (2014)), y d) imágenes recogidas para la elaboración de la presente publicación y las diversas pruebas realizadas. Como imágenes negativas, se capturaron un total de 1000 imágenes desde el interior de cinco vehículos diferentes (200 imágenes de cada uno) para obtener cierta variabilidad, y que el clasificador no clasificara características irrelevantes.…”

Deteksi Kantuk Pengendara Roda Empat Menggunakan Haar Cascade Classifier Dan Convolutional Neural Network

“…Mata merupakan salah satu indra manusia yang digunakan untuk pendeteksian reaksi manusia oleh perangkat cerdas. Analisis ukuran permukaan iris dapat digunakan untuk menentukan keadaan mata dalam kondisi terbuka ataupun tertutup [2]; Beberapa penelitian pendeteksi kantuk menggunakan pengolahan citra digital pernah dilakukan [3], [4] dan [5].…”