RICH AND EFFICIENT VISUAL DATA REPRESENTATION

Increasing the size of training data in many computer vision tasks has shown to be very effective. Using large scale image datasets (e.g. ImageNet) with simple learning techniques (e.g. linear classifiers) one can achieve state-of-the-art performance in object recognition compared to sophisticated learning techniques on smaller image sets. Semantic search on visual data has become very popular. There are billions of images on the internet and the number is increasing every day. Dealing with large scale image sets is intense per se. They take a significant amount of memory that makes it impossible to process the images with complex algorithms on single CPU machines. Finding an efficient image representation can be a key to attack this problem. A representation being efficient is not enough for image understanding. It should be comprehensive and rich in carrying semantic information. In this proposal we develop an approach to computing binary codes that provide a rich and efficient image representation. We demonstrate several tasks in which binary features can be very effective. We show how binary features can speed up large scale image classification. We present learning techniques to learn the binary features from supervised image set (With different types of semantic supervision; class labels, textual descriptions). We propose several problems that are very important in finding and using efficient image representation.

Sparse Signal Representation in Digital and Biological Systems

Theories of sparse signal representation, wherein a signal is decomposed as the sum of a small number of constituent elements, play increasing roles in both mathematical signal processing and neuroscience. This happens despite the differences between signal models in the two domains. After reviewing preliminary material on sparse signal models, I use work on compressed sensing for the electron tomography of biological structures as a target for exploring the efficacy of sparse signal reconstruction in a challenging application domain. My research in this area addresses a topic of keen interest to the biological microscopy community, and has resulted in the development of tomographic reconstruction software which is competitive with the state of the art in its field. Moving from the linear signal domain into the nonlinear dynamics of neural encoding, I explain the sparse coding hypothesis in neuroscience and its relationship with olfaction in locusts. I implement a numerical ODE model of the activity of neural populations responsible for sparse odor coding in locusts as part of a project involving offset spiking in the Kenyon cells. I also explain the validation procedures we have devised to help assess the model's similarity to the biology. The thesis concludes with the development of a new, simplified model of locust olfactory network activity, which seeks with some success to explain statistical properties of the sparse coding processes carried out in the network.

Interaction of Acoustic Waves with a Laminar Line-Flame

A systematic study was conducted to elucidate the effects of acoustic perturbations on laminar diffusion line-flames and the conditions required to cause acoustically-driven extinction. Flames were produced from the fuels n-pentane, n-hexane, n-heptane, n-octane, and JP-8, using fuel-laden wicks. The wicks were housed inside of a burner whose geometry produced flames that approximated a two dimensional flame sheet. The acoustics utilized ranged in frequency between 30-50 Hz and acoustic pressures between 5-50 Pa. The unperturbed mass loss rate and flame height of the alkanes were studied, and they were found to scale in a linear manner consistent with Burke-Schumann. The mass loss rate of hexane-fueled flames experiencing acoustic perturbations was then studied. It was found that the strongest influence on the mass loss rate was the magnitude of oscillatory air movement experienced by the flame. Finally, acoustic perturbations were imposed on flames using all fuels to determine acoustic extinction criterion. Using the data collected, a model was developed which characterized the acoustic conditions required to cause flame extinction. The model was based on the ratio of an acoustic Nusselt Number to the Spalding B Number of the fuel, and it was found that at the minimum speaker power required to cause extinction this ratio was a constant. Furthermore, it was found that at conditions where the ratio was below this constant, a flame could still exist; at conditions where the ratio was greater than or equal to this constant, flame extinction always occurred.