Eye gaze tracking

29
EYE GAZE TRACKING FOR HUMAN COMPUTER INTERFACE Submitted by Prajnyadeep Subhashis Patra Dibyesh Das Priyatosh Das Alok Ahuti Roul Guided by Prof. Bibhu Prasad Mohanty

Transcript of Eye gaze tracking

Page 1: Eye gaze tracking

EYE GAZE TRACKING FOR HUMAN COMPUTER INTERFACE

Submitted by

Prajnyadeep Subhashis Patra

Dibyesh Das

Priyatosh Das

Alok Ahuti Roul

Guided by

Prof. Bibhu Prasad Mohanty

Page 2: Eye gaze tracking

EYE GAZE TRACKING

What is it? Gaze tracking, as the name suggest, is tracking the gaze of a person.

The objective of a gaze tracking system is to track the focus of the user's vision.

So it is necessary to be able to track the eye first from which other information can be inferred.

Page 3: Eye gaze tracking

IS GAZE TRACKING USEFUL?Gaze tracking has a number of possible applications that is beneficial for the society.

A new form of Human-Computer Interaction, that can help aiding the disabled.

The technique can be used to detect when a driver is getting drowsy.

It can be a tool for study of user’s visual behaviour when surfing the internet or for cognitive analysis.

It finds application in medical research and for surveillance purpose.

Recent application of Eye tracking involves the scrolling effect like when we move our eye up or down accordingly it will scroll up or down.

Page 4: Eye gaze tracking

HUMAN VISUAL SYSTEM

The parts of the eye which are of interest for eye-trackers.

Page 5: Eye gaze tracking

IMAGE FORMATION Cornea lets the light inside of the eye, which passes through pupil.

Around pupil, we find the iris. Iris controls the amount of light which passes through pupil.

The light that enters the eye is refracted by the eye lens, and then it falls onto the retina layer which covers the inner surface of the eyeball, forming the image of what is seen through the eyes.

Page 6: Eye gaze tracking

EYE MOVEMENTS

The eye movements can be mainly classified into three categories.

Saccades:

These are rapid eye movements that are aimed at focusing the visual attention to a different object.

Smooth pursuits:These happen when the eyes are following a moving target. In

order to achieve this movement, the speed of the target should be in the range that can be handled by the human eye.

Fixations:These are the movements which stabilize the retina on the object of

interest.

Page 7: Eye gaze tracking

DIFFERENT EYE GAZING TECHNIQUES

Electro-Oculography

Sceleral Search Coils

Infared Oculography

Video Oculography

Page 8: Eye gaze tracking

WORK DONE SO FAR

Our objective is to track the gaze of an eye.

We have been able to detect the eye for 3 directions that is left,middle and right.

Here we have worked with still image.

The complete workflow is as follows:

Page 9: Eye gaze tracking

ENTIRE PROCESS DONE

Input ImageImage

EnhancementImage

Segmentation

BinarizationMorphological

OpeningImage

Cropping

Iris Detection

Page 10: Eye gaze tracking

IMAGE ENHANCEMENT

The principal objective of enhancement is to process an image so that the result is more suitable than the original image.

Here image enhancement is done using laplacian operator.

Page 11: Eye gaze tracking

LAPLACIAN OPERATOR

The Laplacian operator is an example of a second order or second derivative method of enhancement.

It is particularly good at finding the fine detail in an image.

This Operator can be used to find edges in an image.

It can be of two types:

Positive laplacian operator

Negative Laplacian Operator

Page 12: Eye gaze tracking

0 1 0

1 -4 1

0 1 0

0 -1 0

-1 4 -1

0 -1 0

2. Negative Laplacian Operator

1. Positive laplacian operator

Page 13: Eye gaze tracking

INPUT IMAGE

Page 14: Eye gaze tracking

ENHANCEMENT RESULT

(LAPLACIAN MASK RESULT) (ENHANCED IMAGE)

Page 15: Eye gaze tracking

IMAGE SEGMENTATION

Image segmentation is the process of partitioning a digital image into multiple segments.

More precisely, image segmentation is the process of assigning a label to every pixel in an image such that pixels with the same label share certain characteristics.

Here the image segmentation is done on the enhanced image using k-means clustering.

Page 16: Eye gaze tracking

K-MEANS CLUSTERING

Clustering is the process of partitioning a group of data points into a small number of clusters.

The k-means algorithm is an evolutionary algorithm that gains its name from its method of operation.

The algorithm involves clustering of observations into k groups, where k is provided as an input parameter.

The steps for K-mean clustering is as follows.

Page 17: Eye gaze tracking

Take input value k

start

Select centroid for each cluster

Calculate the euclidean distance between each pixel and cluster

centroid

Group the pixel with the cluster having minimum distance

If the new observation changes

the cluster ?

END

Yes

No

FLOW CHART FOR K-MEANS CLUSTERING

Page 18: Eye gaze tracking

OUTPUT OF KMEAN CLUSTERING

Page 19: Eye gaze tracking

BINARIZATION

After segmentation we go for binarization of segmented image on which morphological operation will be done.

Image binarization converts an image of up to 256 gray levels to a black and white image.

The simplest way to use image binarization is to choose a threshold value, and classify all pixels with values above this threshold as white, and all other pixels as black.

Page 20: Eye gaze tracking

BINARIZATION OUTPUT

Page 21: Eye gaze tracking

MORPHOLOGICAL OPENING

In binarization, binary regions produced by simple thresholding are distorted by noise and texture.

Morphological image processing pursues the goals of removing these imperfections by accounting for the form and structure of the image.

Morphological Opening is same as ‘erosion followed by dilation’.

Opening removes small objects from the foreground (usually taken as the dark pixels) of an image, placing them in the background

Page 22: Eye gaze tracking

OUTPUT AFTER MORPHOLOGICAL OPENING

Page 23: Eye gaze tracking

IMAGE CROPPING

The image was cropped in order to find the correct iris position in both eyes.

The original image was cropped into two images-one of left eye and one of right eye.

Page 24: Eye gaze tracking

IRIS POSITION DETECTIONThere were 8 parameters calculated, namely: (min_x, y_min_x,max_x, y_max_x, min_y, x_min_y, max_y,x_max_y).

After calculating the coordinates of the eye, a variable by the name of loc was declared to indicate the direction the user was looking.

The loc variable could take on one of three values at a single time:-

IRIS POSITION LOC VALUE

left 1

right 2

center 3

Page 25: Eye gaze tracking

(Eye looking towards center)

(Eye looking towards right)

(Eye looking towards left)

Page 26: Eye gaze tracking

FLOW CHART FOR IRIS POSITION DETECTION Input

cropped image

count the number of 0’s and 1’s in the region enclosing the area from min_y to max_y

If no of black

pixels >no of white pixels

loc is set to 3.iris is at

center

count white pixels(w1,w2) in region 1 and 2

Region 1:-max_y to min_x

Region2:-max_yand max_x

If W1<W2

Loc=1.Iris is

towards left

Loc=2.Iris is towards

right

YesNo

Yes No

Page 27: Eye gaze tracking

WORKFLOW FOR 8TH SEMTill now we have been able to detect the Iris position in 3 different direction that is right,left and center. Further we will try to detect the Iris position in all possible direction.

And here we have performed the whole process on a single still image.At a later stage we will implement it for continuous images or videos .

The outputs obtained here are not accurate enough however outputs can be improved by finding better techniques through technique comparisons.

Then our ultimate aim is to implement this ideas for

developing a ‘Human-Computer Interface’.

Page 28: Eye gaze tracking

REFERNCES

[1] Vision-Based Eye-Gaze Tracking for Human Computer InterfaceKyung-Nam Kim* and R. S. Ramakrishna*

[2] A SURVEY ON EYE-GAZE TRACKING TECHNIQUESR. Chennamma

[3] Gaze Tracking Using A Regular Web Camera Submitted in partialfulfillment of the requirements of the degree of Bachelor of Science(Honours) of Rhodes University .David Jonathan Wild

[4] Design and Impletation of A HumanComputer Interface Tracking System Based on Multiple EYE FEATURES

1Shazia Azam, 2Aihab Khan, 3M.S.H.Khiyal

[5] Digital image processingRafael C Gonzalez, Richard E Woods

Page 29: Eye gaze tracking

Thank You