National Institute of Technology, Calicut, Kerala, India
Ph.D - Electronics & Communication Engineering
Area of Work: Pattern Recognition, Face Pose Estimation
Month & Year of Passing: May 2019
Thesis: Clustered Manifolds for Face Pose Estimation and Driver Distraction Monitoring
Face pose estimation is a crucial problem in the field of computer vision and robotics. In a computer vision context, face pose estimation is the process of inferring the orientation of a human head from digital imagery. Human computer interface, 3D face modelling, gaze identification, facial expression recognition and face recognition are some of the general major research areas under pose estimation tasks. For analysing any face related problem in this field, face pose angle estimation is a major task. The pose direction can also provide useful information for the inference of person's intent and behaviour. In my research, multi-layer discriminant analysis is used for pose estimation. In this work, multiple projection planes are used for classification.
One of the applications of estimated head pose is driver distraction monitoring. This application concentrates on the development of an automated system that would alert the driver of a vehicle to distractions of self. The number of deaths and injuries due to the motor vehicle accidents have increased drastically during the last few years. It is estimated that nearly 40 - 50 % of the accidents are due to the distractions of the driver. An automated system in the vehicle that would alert the driver of the vehicle to distractions of self will drastically reduce the number of accidents due to this problem. This work monitors the distractions of a driver based on the estimated face pose information. The major problem of driver distraction monitoring is the unavailability of a standard database for testing. Hence, a new database is developed exclusively for driver distraction monitoring and it is used for experimental analysis.
Amrita Vishwa Vidyapeetham University, Coimbatore, Tamilnadu, India
M.Tech.- Computer Vision and Image Processing
Month & Year of Passing: July 2009
CGPA: 7.71 out of 10
Thesis: Evaluation of Different Line Detection Algorithms in SONAR Images
This work completed from Naval Physical and Oceanographic Laboratory (NPOL), Defence Research and Development Organization (DRDO), Ministry of Defence, Government of India, Thrikkakkara, Cochin, Kerala, India.
The proposed work mainly aims the detection of Lines from SONAR images. The normalization and image enhancement are some pre-processing techniques for noise removal before the line detection. Knowledge about lines of a SONAR image is useful in the case of target parameter extraction. The lines will give the parameters like speed, position, path of the target.
Mid-Point Hough Transform algorithm is a new proposed Hough Transform algorithm in this work; which is a modification of Hough Transform based on the geometric properties of parallel lines and different Randomized Hough Transform algorithms. This work also aims to evaluate the performance of different line detection algorithms based on the number of transactions from image space to parameter space, execution time and accuracy.
MES College of Engineering, Kuttippuram, Kerala, India
(Affiliated to University of Calicut)
B.Tech - Applied Electronics & Instrumentation
Month & Year of Passing: June 2007
Marks: 66.9 %
Thesis: Intelligent Crawler
Intelligent Crawler is a six legged robot. It has the ability to walk forward, backward and turn to sides. The design of 12 stepper motors to control the leg movements – a pair for each leg. PIC 16F877 is the main controller used for coordinating its actions. This controller generates the control signal responsible for the leg movements and sensor feedback. PIC 16F84 is used for generating driver signals for the stepper motors based on the control signals by PIC 16F877.
Mini Project: Digital Speedometer
Digital Speedometer is designed for the purpose of numerically displaying the speed of a vehicle in km/hr. Opto-coupling method is incorporated in our circuit contrary to the ordinary magnetic coupling method.