Brian Temu

Data Science Major & Software Engineer

Vision Transformer

Paper Replication

A vision transformer, or ViT, is a deep learning architecture designed for computer vision tasks. Unlike traditional convolutional neural networks (CNNs), which rely on convolutional layers to process image data, ViTs utilize self-attention mechanisms commonly found in transformer models. Vision Transformer Paper ViT extends the success of transformers from natural language processing to image analysis, demonstrating remarkable performance on various image recognition tasks. This introduction will provide an overview of the key concepts and contributions of the Vision Transformer paper, setting the stage for a deeper exploration of its innovative methodology and experimental results.

Real Time Facemask Detection System

Computer Vision

The YOLO Real Face Mask Detection System is an advanced computer vision application that utilizes the YOLO (You Only Look Once) object detection algorithm to identify and categorize whether individuals in images or video streams are wearing face masks accurately and in real-time. This system plays a crucial role in monitoring and enforcing mask-wearing compliance, especially during health crises such as the COVID-19 pandemic. It provides rapid and reliable detection of individuals wearing masks and those who are not, contributing to public health and safety efforts by minimizing the spread of infectious diseases.

Baltimore Police Department Crime

Data Analysis

Open Baltimore Crime Data Open Baltimore Crime Data Analysis is a data-driven project focused on exploring and interpreting crime data from the city of Baltimore. By leveraging various analytical techniques, this initiative aims to uncover trends, patterns, and insights within the dataset, providing valuable information to local authorities and communities. Through rigorous analysis, it seeks to enhance understanding of crime dynamics, contribute to informed decision-making, and ultimately promote safer and more secure neighborhoods in Baltimore.

Spot the Mask Challenge

Deep Learning

The Spot the Mask Project is an exciting competition-based initiative that harnesses the power of machine learning to classify images exclusively sourced from Zindi. In this competition, participants are tasked with creating and fine-tuning image classification models to discern whether individuals in these images are wearing masks. It represents a significant and innovative challenge within the machine learning community.