Leveraging Multiple Features for Image Retrieval and Matching

The goal of image retrieval and matching is to find and locate object instances in images from a large-scale image database. While visual features are abundant, how to combine them to improve performance by individual features remains a challenging task. In this work, we focus on leveraging multiple features for accurate and efficient image retrieval and matching. We first propose two graph-based approaches to rerank initially retrieved images for generic image retrieval. In the graph, vertices are images while edges are similarities between image pairs. Our first approach employs a mixture Markov model based on a random walk model on multiple graphs to fuse graphs. We introduce a probabilistic model to compute the importance of each feature for graph fusion under a naive Bayesian formulation, which requires statistics of similarities from a manually labeled dataset containing irrelevant images. To reduce human labeling, we further propose a fully unsupervised reranking algorithm based on a submodular objective function that can be efficiently optimized by greedy algorithm. By maximizing an information gain term over the graph, our submodular function favors a subset of database images that are similar to query images and resemble each other. The function also exploits the rank relationships of images from multiple ranked lists obtained by different features. We then study a more well-defined application, person re-identification, where the database contains labeled images of human bodies captured by multiple cameras. Re-identifications from multiple cameras are regarded as related tasks to exploit shared information. We apply a novel multi-task learning algorithm using both low level features and attributes. A low rank attribute embedding is joint learned within the multi-task learning formulation to embed original binary attributes to a continuous attribute space, where incorrect and incomplete attributes are rectified and recovered. To locate objects in images, we design an object detector based on object proposals and deep convolutional neural networks (CNN) in view of the emergence of deep networks. We improve a Fast RCNN framework and investigate two new strategies to detect objects accurately and efficiently: scale-dependent pooling (SDP) and cascaded rejection classifiers (CRC). The SDP improves detection accuracy by exploiting appropriate convolutional features depending on the scale of input object proposals. The CRC effectively utilizes convolutional features and greatly eliminates negative proposals in a cascaded manner, while maintaining a high recall for true objects. The two strategies together improve the detection accuracy and reduce the computational cost.

Event-Code Interaction Directed Test Cases

The Graphical User Interface (GUI) is an integral component of contemporary computer software. A stable and reliable GUI is necessary for correct functioning of software applications. Comprehensive verification of the GUI is a routine part of most software development life-cycles. The input space of a GUI is typically large, making exhaustive verification difficult. GUI defects are often revealed by exercising parts of the GUI that interact with each other. It is challenging for a verification method to drive the GUI into states that might contain defects. In recent years, model-based methods, that target specific GUI interactions, have been developed. These methods create a formal model of the GUI’s input space from specification of the GUI, visible GUI behaviors and static analysis of the GUI’s program-code. GUIs are typically dynamic in nature, whose user-visible state is guided by underlying program-code and dynamic program-state. This research extends existing model-based GUI testing techniques by modelling interactions between the visible GUI of a GUI-based software and its underlying program-code. The new model is able to, efficiently and effectively, test the GUI in ways that were not possible using existing methods. The thesis is this: Long, useful GUI testcases can be created by examining the interactions between the GUI, of a GUI-based application, and its program-code. To explore this thesis, a model-based GUI testing approach is formulated and evaluated. In this approach, program-code level interactions between GUI event handlers will be examined, modelled and deployed for constructing long GUI testcases. These testcases are able to drive the GUI into states that were not possible using existing models. Implementation and evaluation has been conducted using GUITAR, a fully-automated, open-source GUI testing framework.

In Situ Characterization of Optical Absorption by Carbonaceous Aerosols: Calibration and Measurement

Light absorption by aerosols has a great impact on climate change. A Photoacoustic spectrometer (PA) coupled with aerosol-based classification techniques represents an in situ method that can quantify the light absorption by aerosols in a real time, yet significant differences have been reported using this method versus filter based methods or the so-called difference method based upon light extinction and light scattering measurements. This dissertation focuses on developing calibration techniques for instruments used in measuring the light absorption cross section, including both particle diameter measurements by the differential mobility analyzer (DMA) and light absorption measurements by PA. Appropriate reference materials were explored for the calibration/validation of both measurements. The light absorption of carbonaceous aerosols was also investigated to provide fundamental understanding to the absorption mechanism. The first topic of interest in this dissertation is the development of calibration nanoparticles. In this study, bionanoparticles were confirmed to be a promising reference material for particle diameter as well as ion-mobility. Experimentally, bionanoparticles demonstrated outstanding homogeneity in mobility compared to currently used calibration particles. A numerical method was developed to calculate the true distribution and to explain the broadening of measured distribution. The high stability of bionanoparticles was also confirmed. For PA measurement, three aerosol with spherical or near spherical shapes were investigated as possible candidates for a reference standard: C60, copper and silver. Comparisons were made between experimental photoacoustic absorption data with Mie theory calculations. This resulted in the identification of C60 particles with a mobility diameter of 150 nm to 400 nm as an absorbing standard at wavelengths of 405 nm and 660 nm. Copper particles with a mobility diameter of 80 nm to 300 nm are also shown to be a promising reference candidate at wavelength of 405 nm. The second topic of this dissertation focuses on the investigation of light absorption by carbonaceous particles using PA. Optical absorption spectra of size and mass selected laboratory generated aerosols consisting of black carbon (BC), BC with non-absorbing coating (ammonium sulfate and sodium chloride) and BC with a weakly absorbing coating (brown carbon derived from humic acid) were measured across the visible to near-IR (500 nm to 840 nm). The manner in which BC mixed with each coating material was investigated. The absorption enhancement of BC was determined to be wavelength dependent. Optical absorption spectra were also taken for size and mass selected smoldering smoke produced from six types of commonly seen wood in a laboratory scale apparatus.