Anthony Quiroga Email and Phone Number

Worked on a team with thesis advisor Prof. Pat Pannuto (UCSD) and Prof. Jay Taneja (UMass Amherst) that’s focused on creating a cost-effective solution for electric grid quality detection through images of light sources from cameras. Created a camera system to utilizes computer vision techniques, such as stereo vision and imaging geometry, for estimating light sources in 2D images to 3D space in the Global Positioning System (GPS) by using Raspberry Pi systems (Linux OS).Constructed a stereo camera system to handle 3D point estimation of objects at 15m distances within 1m accuracy using custom baseline and applied sensors, such as IMU and GPS tracking, for estimating objects from 2D images to GPS locations. Utilize C++ libraries, such as GeoTrans, Eigen, and OpenCV to conduct the mapping of 2D image points to 3D scene points to estimate light sources in images to GPS coordinates for real-world positioning autonomously.Automated the camera system to detect light sources through color analysis of night-time city lights using OpenCV and Numpy for image processing in Python with 90% True Positive rate.Applied computer vision algorithms, Key Frame detection and Bundle Adjustment, to optimize initial GPS light source position estimates for greater accuracy of real-world positioning of light sources within 12 meters.Utilized the ROS library on Raspbian based camera system to apply Kalman filter optimization on camera orientation estimation for improved GPS localization of light sources by utilizing sensor fusion packages on ROS for IMU, GPS, and Cameras.

Hosted lab hours to assist students with (C++) programming assignments, projects, quizzes that involved deployment or applications of data structures learned in class. Tested course content to ensure assignment accuracy for proper deployment to 500+ students in the course. Answered questions from students that utilized online forum, edstem.org, outside of hosted lab hours.

Designed an algorithm that uses shape recognition and text recognition as computer vision techniques to utilize OCR modules for extracting text from image information for given company products to produce an automated system.Developed code that is maintainable and formatted to industry standard, created unit tests for code to ensure that algorithm is testable and integrated ready to utilize for continuous integration for proper testing practices. Adhered to standard coding guidelines and practices from the company. Learned the standard operating principles and expectation to perform the role of a systems engineer in a multi-disciplinary program.Learned to work in a team environment through the utilization of agile methodologies such as task planning, sprint planning, and team standups for communication on project progress. Specifically, I learned to take in constructive feedback and apply it to work, convey concerns and progress on projects, and work in a fast-paced environment to ensure the project is progressing at the expected rate.

Helped students with weekly labs that apply software engineering principles in short app development projects on android studio focusing on agile process development such as unit testing, GitHub utilization, ZenHub utilization, integration testing, and object-oriented design pattern applications. Ran “customer” demos of students’ course-long project application to verify their team applied agile development practices properly throughout project development, such as proper tests for continuous integration and app functionality, use of SRP and application of design patterns, proper coding practices such as commenting for documentation, and verifying application functionality. Graded for course-long projects, weekly labs, and exams to verify students’ knowledge and proper application of the agile process for software engineering.

Designed automated tests that utilized Raspberry Pi as test harness for Tock embedded systems, expanding continuous integration for the Tock operating system through the utilization of GitHub servers. Worked with embedded systems by successfully creating applications for both Tock supported platforms and the Raspberry PI to test specific Tock platforms, which were included in the Continuous Integration support of Tock Operating System. Authored a paper to convey the project on expanding continuous integration for the Tock platform for the TRELS program, leading to a presentation at the UCSD supported Summer Research Conference for undergraduate students of the TRELS program.

Created a machine learning algorithm, utilizing K-nearest Neighbor, Pandas, SciPy and other Python libraries, to classify and predict interactions between caregiver and child based on proximity data given from the SociTrak devices. Contributed to the SociTrack project, focused on parsing data collected through proximity tracking embedded systems, TotTags, between caregiver and child to develop a system of automating the task of classifying interactions through the data collected. Learned to conduct research under a Professor and an Advisor’s guidance, resulting in learning how to read a research paper, write a research proposal, and present a concise, informative, poster of the conducted research