Detection of eye strain through blink rate and sclera area using raspberry-pi

ABSTRACT

In this paper, a unique approach is proposed to determine eye strain in children caused due to the prolonged exposure to LCD and PDP screens. Blink rate and area of eye sclera are the two efficient metrics relied upon to determine eye strain. Initially, blink rate is computed by monitoring state transitions from open to closed eyes. To segment the sclera region of the eye Otsu’s thresholding and colour tracking, were implemented. These approaches had limitations due to illumination and skin tone. To overcome this, Modified Otsu using Colour Tracking is proposed in this paper. The occurrence of eye strain is determined by comparing the values of the metrics derived in real time with a threshold. The proposed algorithm detects eye strain with an accuracy of 83%. After detection of fatigue an alert message is sent to caretakers.

KEYWORDS:

• Eye fatigue
• blink rate
• sclera
• Otsu
• raspberry-pi

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Alla Phani Charan Reddy received his Bachelors in Electronics and Communication Engineering from VIT Chennai and is currently working as a QA Engineer in Ribbon Communications, India. Research interests include Networking, Embedded systems and Image processing.

B. V. H. Sandilya received his Bachelors in Electronics and Communication Engineering from VIT Chennai. He is working as a Software Engineer in Texas Instruments, India. Research interests include Embedded systems, Image processing and VLSI Design Automation.

A. Annis Fathima obtained BE (ECE) from Madras University, ME (VLSI Design) and Ph.D. in Image processing and pattern recognition from Anna University, Chennai, India. Currently working as Associate Professor, School of Electronics Engineering at VIT Chennai. The area of interest includes Image processing, Video Analytics, Pattern Recognition, Biometrics.

Image notes

Authors confirm that permissions have been obtained for the reproduction of the images contained within figures 3, 5, 9, 11 & 12. The dataset was collected by the co-authors and was prepared with the consent of the students. In case of Figure 3, parental consent has been obtained.