Background Modern cameras are unable to capture the full dynamic range of commonly encountered real-world scenes. In some scenes, even the best possible photograph will be partially under or over-exposed. Researchers and photographers commonly get around this limitation by combining information from multiple exposures of the same scene.

Nov 30, 2024

Part A: The Power of Diffusion Models! Setup For this project, I set the random seed to 180 manually to ensure reproducibility of the results. The DeepFloyd diffusion pipeline utilizes a multi-stage inference process.

Nov 5, 2024

Part A. Image Warping and Mosaicing Taking Images All of these images were taken using an iPhone 15Pro on the street north of campus. Below are some images used in this project.

Oct 18, 2024

Part 1. Defining Correspondences First, I used a labeling tool from a student project last year (labeling tool). I manually added the four corners of an image because this allows for transforming the entire image.

Oct 2, 2024

Part 1: Fun with Filters Finite Difference Operator In this section, I first applied differential filters in the x and y directions separately and visualized them accordingly. For the gradient magnitude, I computed $\sqrt{dx^2+dy^2}$ for each pixel.

Sep 19, 2024

Overview The goal of this project is to take the digitized Prokudin-Gorskii glass plate images and, using modern image processing techniques, automatically produce a color image with as few visual artifacts as possible.

Sep 18, 2024