Collection of selected ECE MEng course projects completed at the University of Ottawa. This repository includes work in deep learning super-resolution, semiconductor packaging thermal analysis, machine vision color detection, and low-power embedded system analysis.
| Folder | Description |
|---|---|
DL_SuperResolution/ |
Deep-learning image super-resolution |
FlipChip_Thermal/ |
Flip-chip packaging thermal analysis (ANSYS) |
MachineVision_Color/ |
Multi-color object detection & position extraction |
MSP430_LowPower_Analysis/ |
Power consumption analysis of MSP430 MCU |
This project implements a frequency-inspired transformer-based super-resolution model. It compares L1 loss and L1 + LPIPS perceptual loss and analyzes their effects on reconstruction quality. Includes training scripts, model architecture, experiment results, and visualization.
A thermal simulation study using ANSYS Mechanical Workbench 2024 R1. The project evaluates the impact of materials, thickness, TIM conductivity, convection coefficient, and power levels on heat dissipation. Results show the lid and TIM dominate thermal performance, with convection being the primary heat transfer mechanism.
A classical machine vision pipeline for multi-color object recognition and position detection. Includes calibration, image preprocessing, segmentation, and object localization. Demonstrates practical applications in industrial detection and robotics.
An analytical and experimental study on MSP430 low-power modes. Examines current consumption under different configurations, clock settings, and sleep modes. Includes report, measurement data, and power-saving optimization analysis.