Computer Vision

Image processing Theory jargon Difference between supervised and unsupervised learning Criteria Supervised Learning Unsupervised Learning Data Uses labeled data for training. Uses unlabeled data for training. Goal Predict a label for new data based on past observations. Discover hidden patterns or intrinsic structures within the data. Examples Classification, Regression Clustering, Association Complexity Less complex as it has a clear goal. More complex due to the lack of clear goal. Usage When the outcome of the problem is known. When the outcome of the problem is unknown. Application #1 Spam Detection Customer Segmentation Application #2 Credit Fraud Detection Anomaly Detection EM spectrum The Electromagnetic Spectrum (EM) is the range of all types of EM radiation. Radiation is energy that travels and spreads out as it goes – visible light that comes from a lamp in your house or radio waves from a radio station are two types of electromagnetic radiation. Other examples of EM radiation are microwaves, infrared and ultraviolet light, X-rays, and gamma-rays. ...

Image Analysis using pytorch Prerequisites This project is built using python in Ubuntu (WSL) and you’ll need to install the following: Any bash terminal (one of the following) Conda WSL Mac OS Any flavour of Linux Python 3 (I’m using 3.10.12) 1 2 sudo apt update sudo apt install python3 Pip 1 2 sudo apt update sudo apt install python3-pip Packages Note: Since I’m using a computer with a CUDA compatable NVIDIA GPU, I’ll be using the GPU version of pytorch. If you don’t have a GPU, you can install the CPU version of pytorch given below. CPU install 1 pip3 install torch torchvision numpy matplotlib GPU install Installing numpy and matplotlib 1 pip3 install numpy matplotlib Installing pytorch Check the pytorch website to see the which library is compatable with your system. In my case I’m using CUDA 11.8 1 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 Jupyter notebook (optional) 1 pip3 install jupyterlab Or just use Jupyter notebook from VS Code from here Environement I’ve used 1 CPU Ryzen 7 5800H 2 GPU RTX 3060 Laptop 3 RAM 2x8GB DDR4 @ 3200MHz 4 OS Windows 11/Ubuntu 22.4 WSL 5 CUDA 11.8 6 Python 3.10.12 MNIST number dataset The MNIST dataset is a dataset of handwritten digits. It has 60,000 training images and 10,000 test images. We’ll see a code to load the dataset and display the occurances of individual digits in the dataset. Or if you want to run the code from Jupyter notebook you can clone my repository via git. ...