I am a graduate student in the Electrical Engeineering department at North Carolina State University, where my areas of concentration include Pattern Recognition, Machine Learning and Computer Vision. In the past, I have worked at Vodafone, Central Water & Power Research Station, Gov. of India and going to join Nokia - Alcatel Lucent in January 2017. I have a Bachelor's degree in Electronics & Telecommunciation Engineering from Pune Institute of Computer Technology.

I play the Indian Classical Harmonium. I have been playing this beautiful instrument since childhood and have accompanied many established musicians over the years. Recently I have taken to the genre of Progressive Rock Music on the recommendation of a friend and have really started to appreciate it. Do checkout some videos and recordings here.

On the web, you can find me on LinkedIn, Facebook, Github, and Soundcloud. The best way to contact me is over email: ppprabhu@ncsu.edu.

CRF for Image Segmentation

Modeled Foraminifera images as Conditional Random Fields and most probable labeling is achieved using engergy minimization (node and edge potentials) by Hammersley-Clifford Theorem. The non-convex optimization problem is solved using Mean Field Annealing and Linear Programming and Iterated Conditional Modes are used for label prediction. ( report )

Expertiza (Open Source)

Expertiza is a Ruby on Rails web application where students can submit and peer-review learning objects (articles, code, web sites, etc). I was involved in the design and development of the topic-team assignment module. I even worked on redesigning the user interface for assignment creation and editing using ReactJS. This project is sponsored by National Science Foundation. ( wiki1 | wiki2 )

Room Reservation System for NCSU Libraries

Designed and developed a Room Reservation System for NCSU Libraries using Ruby on Rails MVC architecture. The system is equipped with user (role) based authentication, quick glance views for admins and superadmins, reservation histories, notification mailers, bootstrap UI etc. The project was developed under the Object Oriented Design and Development course. ( Demo | Source )

Daytime Water Segmentation

This project focused on segmentation and classification of the water bodies using various machine learning algorithm such as Perceptron (Punish and Reward), Bayesian, Neural Nets, Suppport Vector Machine, Clustering, LDA, PCA etc. Developed Adaptive Sky Removal which removes sky recursively. ( code | report )

VCD for edge preserving smoothing

Edge preserving smoothing of an image is achieved using Variable Conductance Diffusion (Green's function). Project involved formulating the PDE and solving that PDE practically to achieve the optima. ( code | report )

Open Stream Water Velocity Profiling

Research was carried out at Central Water and Power Research Station, India for the Navi Mumbai International Airport safety. The concept of Particle Image Velocimetry is used for the velocity profiling. This research successfully eliminated the use of VADCP (Verticle Acoustic Doppler Current Profiler) sensors and reduced the cost by 90%. ( report )

Biobot Activity Recognition

Activity Recognition and prediction is performed for biobots (in this case cockroaches mounted with gyroscope, accelerometer and other sensors) using Hidden Markov Models. Fine Nearest Neighbor is used for prediction of the states in the first stage and this prediction is used as a feature input the HMM to calculate transition and emission matrices which in turn are used for estimation. ( report )

Deception

It's a game developed in JavaScript and WebGL at NC State as a capstone project. Third party libraries such as Three.js and Cannon.js are used in the development for 3D modeling and physics. The game is inspired from the old SkyRoads and features story mode, background score, SFX, 3D views, physics etc.
( Play the Game! | source )

Raycasting, Rasterization, Textures

These small demos are developed during the Computer Graphics course at NC State using JavaScript, WebGL and GLSL (OpenGL Shading Language).
Raycasting ( Demo | Source )
Rasterization ( Demo | Source )
Textures ( Demo | Source )

Panoramic Image Stitching

Designed and developed the scale, rotation and illumination invarient image matching and stitching using SIFT (Scale Invarient Feature Transform) descriptor. SIFT is used for the keypoint localization and obtaining the minimum distance matching points from both the images. Principle of homography is used for stitching. ( code | report )

Circular Extension to Hough Transform

Hough Transform is modified to detect the circular shapes in an image. It is implemented using 3D accumulator as well as the gradient based method. Multiple circle detection is achieved using k-means clustering. ( code | report )

Sedimentation: Classification and Quantization

Research was carried out at Central Water and Power Research Station, India for the efficient water management in dams. This research focused on increasing dead water storage capacity and effective scheduling of the dam maintenance activities. Adaptive Color Gamma Correction is used for the classification and quantization purposes.

Prathamesh Prabhudesai, Saif Sayed, M Selva Balan, Dr. Sushma Shelke, Aug 2012, Quantization and Classification of Sedimentation of a Coastal Hydraulic Model using Image Processing Technique, International Journal for Research and Review in Applicable Mathematics and Computer Science, Vol. 3, No. 1

Prathamesh Prabhudesai, Saif Sayed, M Selva Balan, Dr. Sushma Shelke, Dec 2012, Adaptive Color Gamma Correction For Sediment Classification and Quantization in a Reservoir, National conference on Hydraulics and Water Resources, HYDRO 2012, Indian Institute of Technology, Bombay, Mumbai, IN

Prathamesh Prabhudesai, Saif Sayed, M Selva Balan, Dr. Sushma Shelke, July - August 2014, A Novel Method For Measurement Of An Open-Stream Water Flow For A Hydrological Model With Multiple Tracers Using Image Processing, International Journal of Science and Applied Information Technology, Vol. 3, No. 4

Please download a copy of my resume from here.