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.

The Effect of Psychological Distance on Willingness to Engage in Ideologically Based Violence

Research on the cognitive and decision-making processes of individuals who choose to engage in ideologically based violence is vital. Our research examines how abstract and concrete construal mindsets affect likelihood to engage in ideologically based violence. Construal Level Theory (CLT) states that an abstract mindset (high-level construal), as opposed to a concrete mindset (low-level construal), is associated with a greater likelihood of engaging in goal-oriented, value-motivated behaviors. Assuming that ideologically based violence is goal-oriented, we hypothesized that high-level construal should result in an increased likelihood of engaging in ideologically based violence. In the pilot study we developed and tested 24 vignettes covering controversial topics and assessed them on features such as relatability, emotional impact, and capacity to elicit a violent reaction. The ten most impactful vignettes were selected for use in the primary investigations. The two primary investigations examined the effect of high- and low-level construal manipulations on self-reported likelihood of engaging in ideologically based violence. Self-reported willingness was measured through an ideological violence assessment. Data trends implied that participants were engaged in the study, as they reported a higher willingness to engage in ideologically based violence when they had a higher passion for the vignette's social issue topic. Our results did not indicate a significant relationship between construal manipulations and level of passion for a topic.