Unleashing Visual Intelligence
Dive into the mesmerizing world of computer vision! Explore my groundbreaking projects that transform pixels into powerful insights. Join me on this visual adventure where technology meets creativity, and let’s redefine what’s possible together!
I am a Computer Vision and AI enthusiast working on 3D reconstruction and Visual Language Models. With both technical and research experience in diverse areas of AI,CV, Robotics and Data Science, and strong communication and public talking skills, I am passionate about using AI to solve challenging, impactful problems.
In my free time, I enjoy basketball, swimming, kickboxing, and teaching AI to youth in the community.
Education
Major Coursework:
Deep Learning and Fuzzy logic, Digital Image Processing, Speech Processing, Deep learning for Autonomous Vehicles, Basics of Machine Learning, Business Analytics, Data Structures using C, Control System Engineering, Signals and Systems, Digital Signal Processing, Audio Video Engineering, Database Management Systems and Embedded Systems Design.
Worked on Computer Vision, Deep Learning, Machine Learning, Data Science and Robotics!
01
● Image processing, Object detection, Tracking
● Point Cloud segmentation, 3D Reconstruction, Deep learning.
● Pytorch, OpenCv, Open3D, Tensorflow, GANs
● Stable Diffusion, YOLO, OCR
● Python, MATLAB and C++
02
● Mathematics and Statistics
● NumPy, Pandas, Sci-kit
● Seaborn, Matplotlib, Sklearn,
● Machine learning models like regression, classification and clustering algorithms
● EDA
● Speech Processing in Python and MATLAB
03
● Docker
● MongoDB
● Mqtt
● Flask
● Steamlit
● Google Cloud
04
● Microcontrollers: Arduino, ESP32 and STM32
● Embedded computing: RaspberryPi and Jetson
● Headless display configuration and automating systems
● Various electronics sensors.
3D Mapping System
Leveraging Autonomous Driving Technology to solve for Air Pollution.
Compute in a hardshell backpack.
Simple intuitive UI for non-technical users.
Point Cloud Analysis
Implemented shapefitting algorithms for pointcloud validation against 3D CAD models
Researched and developed algorithms for automated 3D shape reconstruction from pointcloud data
Developed C++ implementations for point cloud preprocessing and segmentation
Spatial Insights and Room Specific suggestions
Worked on extensive trend analysis on pollution data from indoor air purifiers and provided customized suggestions and graphs for business and health insights.
Vision oriented analysis
Developed a time-stamp based footfall counter analysis tool for a specific customer
Entry-Exit counter
Developed a CCTV-based monitoring platform with advanced applications for workplace surveillance, implementing features such as entry/exit tracking and machine-specific personnel counting, significantly enhancing facility security and operational efficiency. Tools used: Flask, OpenCv, DeepSort, Pytorch and Yolo.
Automatic Exploratory analysis
Engineered a customizable automatic comprehensive data filtration, resampling, and KPI dashboard that generates reports on features and data characteristics from the IoSense platform, enhancing data insights and decision-making capabilities.
Member of Propulsion and Avionics Team (October 2021- June 2022)
● Developed a Real-Time Autonomous Target Identification System utilising Image Processing on a Raspberry Pi to detect the landing zone and retrieve the GPS coordinates for the Powered Autonomous Delivery Aircraft to land on target.
● Built an Electrical Propulsion System combining BLDC motor, ESC, propeller, and battery of a Fixed-wing Advanced Class RC Aircraft and an Arduino Nano-based Embedded Flight Controller System to autonomously control.
● At the SAE Aero Design East competition in May 2022 in Texas, U.S.A., Our team was placed 6th worldwide, 2nd in Asia-Pacific, and 4th in Best Design Report.
Exploring the Applications of Computer Vision and Deep Learning in Various Industries and Use Cases!
IEEE Dataset and Preprint
1. Implemented a novel approach leveraging stable diffusion and image processing to significantly augment a malaria parasite image dataset.
2. This resulted in a near two-fold increase in images per class, providing a richer resource for researchers dedicated to improving malaria diagnostics.
Real-time Weed detection using Image Processing and Deep Learning
1. Classified weeds from the crop accurately in real time using YOLO and image processing and segmentation techniques to ensure that excess pesticide is not sprayed on the crop and weeds are successfully eliminated.
2. Oversaw the project's research, training, and image-processing aspects as the team leader.
3. Enhanced model performance from 85% to 93% accuracy by implementing an image processing layer and expanding the dataset.
Devised a system using a pre-trained CNN model to classify six emotions and tested it using a live camera feed.
1. Worked on a camera-based system to identify objects and guide the blind person to avoid obstacles using Computer Vision tools such as YOLO, PyTorch, OpenCV, and Text to Speech.
2. Researched on various techniques to convert and process 360 degree panoramas and perform object detection.
3. Performed in an interdepartmental team to collect datasets, research, and find effective solutions.
Let’s Connect and Collaborate!