Abstract

The problem of ranking a set of visual samples according to their relevance to a query plays an important role in computer vision. The traditional approach for ranking is to train a binary classifier such as a support vector machine (svm). Binary classifiers suffer from two main deficiencies: (i) they do not optimize a ranking-based loss function, for example, the average precision (ap) loss; and (ii) they cannot incorporate high-order information such as the a priori correlation between the relevance of two visual samples (for example, two persons in the same image tend to perform the same action). We propose two novel learning formulations that allow us to incorporate high-order information for ranking. The first framework, called high-order binary svm (hob-svm), allows for a structured input. The parameters of hob-svm are learned by minimizing a convex upper bound on a surrogate 0-1 loss function. In order to obtain the ranking of the samples that form the structured input, hobsvm sorts the samples according to their max-marginals. The second framework, called high-order average precision svm (hoap-svm), also allows for a structured input and uses the same ranking criterion. However, in contrast to hob-svm, the parameters of hoap-svm are learned by minimizing a difference-of-convex upper bound on the ap loss. Using a standard, publicly available dataset for the challenging problem of action classification, we show that both hob-svm and hoap-svm outperform the baselines that ignore high-order information.

Abstract

Real world applicability of many computer vision solutions is constrained by the mismatch between the training and test domains. This mismatch might arise because of factors such as change in pose, lighting conditions, quality of imaging devices, intra-class variations inherent in object categories etc. In this work, we present a dictionary learning based approach to tackle the problem of domain mismatch. In our approach, we jointly learn dictionaries for the source and the target domains. The dictionaries are partially shared, i.e. some elements are common across both the dictionaries. These shared elements can represent the information which is common across both the domains. The dictionaries also have some elements to represent the domain specific information. Using these dictionaries, we separate the domain specific information and the information which is common across the domains. We use the latter for training cross-domain classifiers i.e., we build classifiers that work well on a new target domain while using labeled examples only in the source domain. We conduct cross-domain object recognition experiments on popular benchmark datasets and show improvement in results over the existing state of art domain adaptation approaches.

Abstract

In this paper we propose a framework for recognition and retrieval tasks in the context of scene text images. In contrast to many of the recent works, we focus on the case where an image-specific list of words, known as the small lexicon setting, is unavailable. We present a conditional random field model defined on potential character locations and the interactions between them. Observing that the interaction potentials computed in the large lexicon setting are less effective than in the case of a small lexicon, we propose an iterative method, which alternates between finding the most likely solution and refining the interaction potentials. We evaluate our method on public datasets and show that it improves over baseline and state-of-the-art approaches. For example, we obtain nearly 15% improvement in recognition accuracy and precision for our retrieval task over baseline methods on the IIIT-5K word dataset, with a large lexicon containing 0.5 million words.

Abstract

In this paper, we present a solution to generate semantically richer descriptions and instructions for driver assistance and safety. Our solution builds upon a set of computer vision and machine learning modules. We start with low-level image processing and finally generate high-level descriptions. We do this by combining the results of the image pattern recognition module with the prior knowledge on traffic rules and larger context present in the video sequence. For recognition of road markings, we use a SVM based classifier and HOG based classifier. We test our method on real data captured in urban settings, and report impressive performance. Qualitative and quantitative performance of various modules are presented.

Abstract

Proteins display characteristic dynamical signatures that appear to be universal across all proteins regardless of topology and size. Here, we systematically characterize the universal features of fast side chain motions in proteins by examining the conformational energy surfaces of individual residues obtained using enhanced sampling molecular dynamics simulation (618 free energy surfaces obtained from 0.94 μs MD simulation). The side chain conformational free energy surfaces obtained using the adaptive biasing force (ABF) method for a set of eight proteins with different molecular weights and secondary structures are used to determine the methyl axial NMR order parameters (Oaxis 2), populations of side chain rotamer states (ρ), conformational entropies (Sconf), probability fluxes, and activation energies for side chain interrotameric transitions. The free energy barriers separating side chain rotamer states range from 0.3 to 12 kcal/mol in all proteins and follow a trimodal distribution with an intense peak at∼ 5 kcal/mol and two shoulders at∼ 3 and∼ 7.5 kcal/mol, indicating that some barriers are more favored than others by proteins to maintain a balance between their conformational stability and flexibility. The origin and the influences of the trimodal barrier distribution on the distribution of Oaxis 2 and the side chain conformational entropy are discussed. A hierarchical grading of rotamer states based on the conformational free energy barriers, entropy, and probability flux reveals three distinct classes of side chains in proteins. A unique nonlinear correlation is established between Oaxis 2 and the side chain rotamer populations (ρ). The …

Abstract

Entity Disambiguation is the task of associating entity name mentions in text to the correct referent entities in the knowledge base, with the goal of understanding and extracting useful information from the document. Entity disambiguation is a critical component of systems designed to harness information shared by users on microblogging sites like Twitter. However, noise and lack of context in tweets makes disambiguation a difficult task. In this paper, we describe an Entity Disambiguation system, EDIUM, which uses User interest Models to disambiguate the entities in the user’s tweets. Our system jointly models the user’s interest scores and the context disambiguation scores, thus compensating the sparse context in the tweets for a given user. We evaluated the system’s entity linking capabilities on tweets from multiple users and showed that improvement can be achieved by combining the user models and the context based models.

Abstract

The last two decades have witnessed an exponential rise in web content from a plethora of domains, which has necessitated the use of domain-specific search engines. Diversity of crawled content is one of the crucial aspects of a domain-specific search engine. To a large extent, diversity is governed by the initial set of seed URLs. Most of the existing approaches rely on manual effort for seed selection. In this work we automate this process using URLs posted on Twitter. We propose an algorithm to get a set of diverse seed URLs from a Twitter URL graph. We compare the performance of our approach against the baseline zero similarity seed selection method and find that our approach beats the baseline by a significant margin.

Abstract

The velocity, volume and variety with which Twitter generates text is increasing exponentially. It is critical to determine latent sub-topics from such tweet data at any given point of time for providing better topic-wise search results relevant to users’ informational needs. The two main challenges in mining sub-topics from tweets in real-time are (1) understanding the semantic and the conceptual representation of the tweets, and (2) the ability to determine when a new sub-topic (or cluster) appears in the tweet stream. We address these challenges by proposing two unsupervised clustering approaches. In the first approach, we generate a semantic space representation for each tweet by keyword expansion and keyphrase identification. In the second approach, we transform each tweet into a conceptual space that represents the latent concepts of the tweet. We empirically show that the proposed methods outperform the state-of-the-art methods.

Abstract

Entity Disambiguation is the task of associating entity name mentions in text to the correct referent entities in the knowledge base, with the goal of understanding and extracting useful information from the document. Entity disambiguation is a critical component of systems designed to harness information shared by users on microblogging sites like Twitter. However, noise and lack of context in tweets makes disambiguation a difficult task. In this paper, we describe an Entity Disambiguation system, EDIUM, which uses User interest Models to disambiguate the entities in the user’s tweets. Our system jointly models the user’s interest scores and the context disambiguation scores, thus compensating the sparse context in the tweets for a given user. We evaluated the system’s entity linking capabilities on tweets from multiple users and showed that improvement can be achieved by combining the user models and the context based models.

Abstract

To determine the significance of urine albumin/creatinine ratio (UACR) as to predict proteinuria in new-onset hypertension and uric acid in women with preeclampsia and its comparison with healthy normotensive pregnant women. Methods. Healthy normotensive pregnant women (n= 45), women with preeclampsia (n= 36) were included in the study. The preeclamptic group was further divided into two subgroups: mild (n= 25) and severe preeclampsia (n= 11). Results. Higher mean values of urine albumin/creatinine ratio (UACR) and uric acid were found in preeclamptic pregnant women (319.05±247.56 mg/g, 5.98±2.1 mg/dl) as compared to healthy normotensive pregnant women (22.15±8.1 mg/g, 4.07±0.91 mg/dl) in the third trimester with highly significant difference (p< 0.001) statistically. When comparison of the mean values of urine albumin/creatinine ratio and uric acid was done between mild (199.9±46.78 mg/g, 5.24±1.89 mg/dl) and severe preeclamptic pregnant women (589.67±305.38 mg/g, 7.64±1.4 mg/dl), the values were found to be elevated in the latter group with highly significant difference (p< 0.001) statistically. The positive and highly significant correlation was found in UACR/blood pressure, uric acid/blood pressure and UACR/uric acid. Conclusion. It was concluded that UACR and serum uric acid levels increased as disease progressed from mild to severe conditions.