Master's Papers and Projects

Permanent URI for this collection

https://hdl.handle.net/10919/86632

Browse

Browsing Master's Papers and Projects by Author "Abbott, A. Lynn"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Comparative Analysis of Facial Affect Detection Algorithms
    Thomas, Ashin Marin (2020-05-22)
    There has been much research on facial affect detection, but many of them fall short on accurately identifying expressions, due to changes in illumination, occlusion, or noise in uncontrolled environments. Also, not much research has been conducted on implementing the algorithms using multiple datasets, varying the size of the dataset and the dimension of each image in the dataset. My ultimate goal is to develop an optimized algorithm that can be used for real-time affect detection of automated vehicles. In this study, I implemented the facial affect detection algorithms with various datasets and conducted a comparative analysis of performance across the algorithms. The algorithms implemented in the study included a Convolutional Neural Network (CNN) in Tensorflow, FaceNet using Transfer Learning, and Capsule Network. Each of these algorithms was trained using the three datasets (FER2013, CK+, and Ohio) to get the predicted results. The Capsule Network showed the best detection accuracy (99.3%) with the CK+dataset. Results are discussed with implications and future work.
  • Thumbnail Image
    Displaying and improving run-length encoded images on a BTOS system
    Geddes, Patrick H. (Virginia Tech, 1992-12-05)
    The goal of this project was decode and display a run-length encoded image, and to apply error concealment techniques to the decoded image. The software was written in the C programming language, a language learned during the course of this project. This software was written to read a run-length encoded ASCII file and translate it to a bit-mapped ASCII file. The software can display the bit-mapped image on a video screen and print it on a laser printer. The software also implements four error concealment techniques and displays the improved image on the video screen and on a laser printer. This software was written for the U.S. Coast Guard's Standard Workstation, running the Unisys Corporation's Burroughs Technology Operating System (BTOS), using the BTOS C compiler. Most of the code should be transportable to other operating systems, but the display functions make use of system library functions, which may not be available on other operating systems.