Pranav Mantini
Education
|
University of Houston, Houston, TX Ph. D Computer Science, Fall 2015 Advisor: Dr. Shishir Shah. Dissertation Topic: Context Based Human Trajectory Forecasting University of Houston, Clear Lake, Houston, TX MS. Computer Science, Spring 2009 Capstone: Using CUDA for Solar Thermal Plant Computation. St. Joseph College of Engg., Affiliated to Anna University, Chennai, India B. Tech. Information Technology, May 2007 Project: Virtual Defense Machine in JAVA uses 1024 bit key. | ||||
Research Interests
|
Computer Vision, Machine Learning, Video Analytics, Software Development, Algorithm Design. Human Motion Analysis and Prediction, 3D Models for Surveillance, Automated Video Surveillance. |
||||
Experience
|
Post Doctoral Fellow Non-functionality and Operational Degradation Detection in Video Surveillance Systems: Surveillance Cameras become non-functional for variety of reasons, from getting simply unplugged to getting struck by lightning. Currently designing a camera fault detection tool to automatically identify such non-functionality and alert the user through an intuitive user interface. Research Assitant 3D Models for Video Surveillance: Designed a complete 3D environment model for simulating video surveillance through networked video surveillance cameras. Human Occupancy Map Estimation: Developed a methodology for estimating the human occupancy map for hallways based on the surrounding 3D geometry. Human Trajectory Forecasting: Developed an algorithm for predicting human motion within indoor environments that incorporates the effect of the surrounding geometry and observed human social norms. Re-Identification using Contextual Features: Implemented an algorithm to improve re-identifiation through incorporating a human motion prediction model to complement existing features with spatio-temporal features. Camera Placement Optimization: Developed an algorithm to locate the best position to setup security cameras based on the human motion in the environment for obtaining effective surveillance. Tracking using Context Based Human Motion Estimation: Developed an algorithm for tracking humans in indoor environments by employing human motion prediction that takes social context and environment into consideration. Teaching Assitant Shared responsibilities for exams, homework assignments, and grades for graduate level course: Digital Image Processing in Fall 2013. Computer Architecture in Spring 2014. Network Administrator/Web Developer, Quantitative Imaging Lab Designed and developed a website for Quantitative Imaging Lab. Network administrator for Windows network and Ubuntu web servers.
Masters Capstone Project Using Cuda for Solar Thermal Computation: Implemented an algorithm for Enabling faster computation of solar thermal energy using NVIDIA CUDA. A complete polygon clipping algorithm was implemented for the computation of the effective area of reflectors that receives sunlight in a solar thermal plant. |
||||
Skills
|
Languages: C/C++, Python, Java, MATLAB. APIs: OpenCV, OpenGL. Tools: Eclipse, Visual Studio, LaTex, Adobe Photoshop. Operating Systems: Unix/Linux, Windows. |
||||
Publications
|
P. Mantini, S.K. Shah. Human Trajectory Forecasting In Indoor Environments Using Geometric Context. P. Mantini, S.K. Shah. Enhancing re-identification through contextual trajectory forecasting. P. Mantini, S.K. Shah. Camera Placement Optimization Conditioned on Human Behavior and 3D Geometry. P. Mantini, S.K. Shah. Multiple People Tracking using Contextual Trajectory Forecasting. P. Mantini, S.K. Shah. A Signal Detection Theory Approach for Camera Tamper Detection. P. Mantini, S.K. Shah. Camera Tamper Detection Techniques for Video Surveillance: A Survey. (Under review) P. Mantini, S.K. Shah. Human Trajectory Prediction within Indoor Scenarios and Its Application in Re-Identification. (Work in progress) |
||||
Patents, Provisionals
|
Optimal Camera Placement for Effective Surveillance, 2015. |
||||
Leadership Activities
|
Lead a three person team of computer scientist and software engineer to interview one hundred customers in five weeks to create a viable bussiness model for a computer vision startup.
Lead a non-profit organization consisting of 20 members. Conducted various fundraising and awareness events for supporting the education of underprivileged children in India. |