Master's Theses

Date of Award


Document Type



Computer Science


Recognition, Multimedia, Digital


"Recognition of digital media is more prevalent in today’s computer culture than ever before. The advent of low cost storage has created a seemingly infinite amount of metadata on the internet, as well as on local machines throughout the world. It is more important than ever to have the capability of quickly and accurately filtering this metadata to find a desired result. Data can be searched using various criteria. For example, text data is searched by analyzing the contents of the text itself. One might execute a search using a method as simple as looking for the ASCII file name, or as complex as parsing large quantities of text and analyzing it with intelligent algorithms. However, searches are not limited to text. Images are also a searchable piece of digital media. Sometimes, an image cannot be searched by file name. Therefore, methods of analyzing images in a digital library are needed to match any input images the user may provide. Both text and image search methods will be discussed throughout this thesis. We will begin with a discussion of the Eigenface algorithm. This algorithm has become an essential area of study for creating more advanced face/feature recognition algorithms. Before Eigenface was created, face recognition was done primarily by pinpointing key features on a face image, such as the eyes, nose, and mouth. However, this process proved to be slow and inefficient. With the creation of Eigenface, the most distinct features in a face image can be stored as Eigenvalues of a matrix that represents the face image. This method was revolutionary for its time, and will be studied in this thesis with both human and animal faces. Our next study will be with Scale-Invariant Feature Transform (SIFT). As the name of the algorithm implies, the result of this process is invariant to scale of size, as well as rotation and noise. Unlike Eigenface, this process does not require a frontal-facing style image for both input images and library images. SIFT provides us with a modern algorithm to compare to Eigenface, giving us the ability to see how the original idea of Eigenface has 1 evolved into a more efficient and effective face detection algorithm. Moving along to text-based analysis, we will explore the idea of Latent Semantic Analysis (LSA). This type of analysis is very important in search engines and various other kinds of semantic analyzers. LSA makes use of matrices and Singular Value Decomposition to sort out the most frequent and important words and phrases in any piece of text. In fact, this procedure allows us to sort through multiple pieces of text to determine which text is most relevant to our search term [6]. This kind of tool has become very powerful in today’s search engines, and is used by the general population every day."



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.