Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.

Integrating Environmental Justice and Social-Ecological Resilience for Successful Adaptation to Climate Change: Lessons from African American Communities on the Eastern Shore of the Chesapeake Bay

This research concerns the conceptual and empirical relationship between environmental justice and social-ecological resilience as it relates to climate change vulnerability and adaptation. Two primary questions guided this work. First, what is the level of resilience and adaptive capacity for social-ecological systems that are characterized by environmental injustice in the face of climate change? And second, what is the role of an environmental justice approach in developing adaptation policies that will promote social-ecological resilience? These questions were investigated in three African American communities that are particularly vulnerable to flooding from sea-level rise on the Eastern Shore of the Chesapeake Bay. Using qualitative and quantitative methods, I found that in all three communities, religious faith and the church, rootedness in the landscape, and race relations were highly salient to community experience. The degree to which these common aspects of the communities have imparted adaptive capacity has changed over time. Importantly, a given social-ecological factor does not have the same effect on vulnerability in all communities; however, in all communities political isolation decreases adaptive capacity and increases vulnerability. This political isolation is at least partly due to procedural injustice, which occurs for a number of interrelated reasons. This research further revealed that while all stakeholders (policymakers, environmentalists, and African American community members) generally agree that justice needs to be increased on the Eastern Shore, stakeholder groups disagree about what a justice approach to adaptation would look like. When brought together at a workshop, however, these stakeholders were able to identify numerous challenges and opportunities for increasing justice. Resilience was assessed by the presence of four resilience factors: living with uncertainty, nurturing diversity, combining different types of knowledge, and creating opportunities for self-organization. Overall, these communities seem to have low resilience; however, there is potential for resilience to increase. Finally, I argue that the use of resilience theory for environmental justice communities is limited by the great breadth and depth of knowledge required to evaluate the state of the social-ecological system, the complexities of simultaneously promoting resilience at both the regional and local scale, and the lack of attention to issues of justice.