Arpita Tugave's Email Addresses & Phone Numbers

Computer Vision Engineer @ Novateur Research Solutions Novateur works with Government and Academics providing solutions in the field on Computer Vision.Worked on a Phase I DOT (US Department of transportation) project under guidance of Dr. Khurram Shafique.• Researched and developed complete pipeline of deep learning based tracking in Tensorflow. Worked extensively on real time object detection and...

Computer Vision Engineer @ Novateur Research Solutions Novateur works with Government and Academics providing solutions in the field on Computer Vision.Worked on a Phase I DOT (US Department of transportation) project under guidance of Dr. Khurram Shafique.• Researched and developed complete pipeline of deep learning based tracking in Tensorflow. Worked extensively on real time object detection and tracking algorithms.• Built Siamese tracking algorithm in TensorFlow Python, adapting region proposal layers and fully connected layers to match objects in consecutive frames. From August 2017 to February 2018 (7 months) VirginiaSoftware Engineering Intern - Machine Learning @ Incode Technologies - ImageIN: Image Editing Social Media platform with the goal to make internet safe. Implemented Image compression for specific requirements of improved quality, reduced image size and bandwidth.- Researched on GANs - Generative adversarial networks to automate image editing platform.- Involved in iOS application development and integration on Objective C and Swift platform. From February 2017 to April 2017 (3 months) San Francisco Bay AreaiOS App Developer @ IBM - Trained in backend technologies of NodeJS, IBM Bluemix, Mobile first platform- Contributed to 3 iOS front end applications From December 2014 to July 2015 (8 months) Bangalore, IndiaComputer Vision Researcher (Volunteer) @ Forus Health Private Limited Working with Dr. Tim Poston, the brain behind marvelous mathematical discoveries, invoked my interests in: thinking in higher dimensional and mathematics as applied to computer vision and graphics. - Ray Caustics: 2D and 3D lens models with approximate ray convergence were designed to understand the idea of imperfect focus. Developed human eye model with different lenses and refractive indices to study focal cusp formation.“Web-interactive, user-friendly publication to help readers better understand the concept of focus – Focal plane and Focal point are a fallacy”- Broader Framework for Lensometry: Developed low cost solution to find the power of the lens using camera, lens, and “mire” image. Reduced software latency using matrices.- Xelular Sets Image Segmentation: Contributed to a novel image segmentation algorithm – improvement over snake active contour technique, captured details of human retina. From December 2013 to July 2015 (1 year 8 months) Bangalore, IndiaSoftware Engineer Intern @ Lennox International Worked as a consultant, helping out with the road map for potential machine learning applications.Analysed data from different time period and region based thermostats. Established deep learning work flow ( using LSTM) which would take normalized data, process and predict the pattern. From May 2017 to July 2017 (3 months) Dallas/Fort Worth AreaGraduate Research Assistant @ University of Florida Dr. Woodard Damon, the director of this lab, guided the team to various successful, ground breaking projects.• Landmine Detection (NSF funded): Implementation of Feature-Extractors such as: GRAB, LBP-variants, HOG, WLD and LPQ in a concurrent Matlab setup. Data pattern visualized for radar signals projected on metallic components and other terrain.• 3D Face Modelling Survey: Face modeling and it's implication in art, social networking, automation explored in detail.• Gender Classification of Iris Bio-metric system: It is surprising that human Iris has gender variations. Biology and other sciences haven't been able to prove this fact. However, we could achieve an overall accuracy of 87% with intelligent modifications on extracted features. Note that: even if the best results achieved so far is 91%, feature extractor identified by our research group far exceeds this in support of lacking dataset. Programmed using Matlab and C++. From September 2015 to December 2016 (1 year 4 months) Bio-Metric and Machine Learning LabSoftware Engineer II - Computer Vision / Graphics @ SMT (SportsMEDIA Technology Corp) - Ongoing development of 3D Modeling application; involves: Architecting from scratch, maintaining JIRA list, weekly meetings to check on requirements and developments.- Develop and Maintain low level camera related libraries. Collaborate with other researchers to keep algorithms updated with that of current research. - Handle problems related to multiple cameras. Development and tweaking on Photometric flow.- Accurate camera position estimation on the fly and camera registration. Exploring improvements of camera tracking with novel approaches. - Usage of Multi View geometry, Bayesian Estimation, GPU computation, OpenGL rendering, Machine learning, and Deep learning techniques.- Software and tools - C++, Python, Qt, OpenGL, Windows, SVN, Linux. San Francisco Bay AreaSoftware Engineer - Perception @ AEye, Inc. Aeye. Inc is a perception sensor company for self-driving industry.- Designed multi-threaded architecture in C++ to handle image and lidar data in parallel.- Implemented Hungarian assignment algorithm and K-Means clustering with smart seeding in CUDA.- Collaborated with other researchers on the fusion of camera and lidar data; extending to 7 degrees of freedom.- Developed perception stack including detection, classification, and tracking (using unscented Kalman filter) of level 3 autonomous driving objects.- Presented Aeye perception capabilities at CES 2019.Software and tools - C++, Python, Qt, Linux From March 2018 to January 2019 (11 months) San Francisco Bay Area