Eye Blink Sensor Assistive Device

Abstract

This study proposes a real-time system that translates voluntary eye blinks into Morse code, enabling communication for individuals with severe motor impairments. Using a standard webcam and computer vision techniques, the system calculates the Eye Aspect Ratio (EAR) to detect intentional blinks, classifies them as short (dot) or long (dash), and decodes these sequences into alphanumeric characters. Five participants completed predefined messaging tasks (“SOS” and “HELP”) using blink sequences, while response time and accuracy were measured. The system achieved an average decoding accuracy of 62%, with response times ranging from 18 to 20 seconds across participants.

These results demonstrate that blink-based Morse code communication is a viable, low-cost alternative to specialized hardware for assistive communication. This approach has potential for enhancing independence and interaction for users with severe motor limitations, particularly in low-resource environments where conventional assistive devices are unavailable or cost-prohibitive.

Introduction

Assistive communication technologies are critical for individuals with conditions that severely limit voluntary movement, such as amyotrophic lateral sclerosis (ALS), spinal cord injury, or locked-in syndrome. Traditional solutions often rely on specialised hardware, such as eye-tracking devices or brain–computer interfaces, which can be expensive and inaccessible in low-resource settings.

This work explores a low-cost, camera-based alternative: translating voluntary eye blinks into Morse code in real time. Morse code, with its binary short/long signal structure, aligns naturally with the human ability to produce two distinct blink durations. By leveraging modern computer vision techniques, we eliminate the need for dedicated sensors, relying only on a standard webcam and open-source software.

Our Contributions Are Threefold:

1. We present a complete pipeline for blink detection, classification, and Morse code decoding that operates in real time using standard hardware. 2. We design and conduct a controlled experimental protocol to evaluate accuracy and response time across multiple participants.

3. We provide quantitative and visual analysis of performance, highlighting the feasibility and limitations of the approach.

Related Work

Blink-based communication systems have a long history, with early implementations relying on dedicated infrared sensors or electromyography (EMG) to detect eyelid muscle activity. While effective, these approaches required specialized hardware and careful calibration, limiting their practical use.

Recent advances in computer vision have enabled non-contact blink detection using standard cameras. Soukupová and ˇCech (2016) introduced the Eye Aspect Ratio (EAR) metric, which calculates the ratio between eyelid landmarks

A Preprint - August 25, 2025

to reliably distinguish open and closed eyes. This technique is computationally efficient and robust to moderate head movement, making it well-suited for real-time applications. Building on EAR-based detection, several systems have been developed to facilitate assistive communication. For example, blink-triggered keyboards and on-screen pointer controls enable users with severe motor impairments to interact with devices. Some methods encode multiple blink patterns to represent distinct commands; however, most focus on discrete control signals rather than continuous text input.

Morse code, a compact and widely recognized binary encoding scheme, has been integrated into assistive devices such as sip-and-puff systems, switch interfaces, and EEG-driven brain-computer interfaces. Although a few studies have combined blink detection with Morse code, these efforts often suffer from limited experimental validation or require expensive hardware.

Our work advances this field by implementing an open-source EAR-based blink detection algorithm combined with real-time Morse code decoding on consumer-grade webcams. This approach aims to provide an accessible, low-cost solution for continuous assistive communication.

The System Operates In Three Sequential Stages:

1. Blink Detection: Mediapipe’s face mesh tracking detects 468 facial landmarks per frame. Specific eye landmarks are used to compute the Eye Aspect Ratio (EAR), a metric introduced by Soukupová and ˇCech (2016), which decreases sharply during eye closure.

2. Blink Classification: EAR values are compared against a calibrated threshold. Blinks are classified as short (dot) or long (dash) based on their duration. 3. Morse Code Decoding: The sequence of dots and dashes is mapped to alphanumeric characters using a predefined Morse code dictionary. Characters are finalized after a pause exceeding the calibrated inter-character time gap.

2 · ∥P1 −P4∥

The numerator represents the sum of vertical eyelid distances, and the denominator represents the horizontal eye width.

Blink Classification Logic

• Short Blink (Dot): Duration ≥1.0 and < 2.0 seconds

• Long Blink (Dash): Duration ≥2.0 Seconds

• Letter Gap: A pause longer than 1.0 second triggers committing the current Morse sequence into a letter. • Word Gap: A pause longer than 3.0 seconds inserts a space, indicating a new word. Thresholds were determined during a brief calibration phase for each participant to account for individual differences in blink speed.

Experimental Protocol

Participants were seated at a fixed distance (approximately 50 cm) from the camera in a well-lit environment. Each participant was instructed to communicate a predefined set of phrases using only voluntary blinks, following Morse code rules.

Trials were logged with the following information:

• Participant Id

• Response time (s) — measured from the start signal to correct message completion • Accuracy — whether the intended message was transmitted correctly • Blink events — timestamps and classification for each blink

Related Journal Articles & DOI Links

Selected peer-reviewed publications relevant to 12 Lead ECG Acquisition. Click the DOI to access the full paper (may require institutional access).

Why Choose Us?

Bangalore guidance for robotics, Spectre and autonomous systems projects.