Abstract. Detection and tracking of carried objects has been of great interest, especially with respect to activity analysis and surveillance. This paper proposes a novel approach for carried object detection and tracking by characterising carried objects given that only the carry event occurs i.e. that these objects follow a person's trajectory with a temporally continuous and characteristically consistent spatial relationship with respect to the person. In addition, we incorporate the use of geometric shape properties such as convexity to detect a generic class of carried objects together with other properties such as track continuity and overlap with protrusions on a person silhouette. We formulate the carried object detection and tracking task as finding the most likely set of tracks associated with a person that has these properties. The proposed approach significantly outperforms a state-of-the-art approach on two challenging datasets PETS2006 and MINDSEYE2012.
Abstract. Mereotopologies have traditionally been defined in terms of the intersection of point sets representing the regions in question. Whilst these semantic schemes work well for purely topological aspects, they do not give any semantic insight into the degree to which the different mereotopological relations hold. This paper explores this idea of a distance based interpretation for mereotopology. By introducing a distance measure between x and y, and for various Boolean combinations of x and y, we show that all the RCC8 relations can be distinguished. We then introduce a distance measure which combines these individual measures which we show reflect different paths through the RCC8 conceptual neighbourhood -i.e. the measure decreases/increases monotonically given certain monotonic transitions (such as one region expanding). There are several possible applications of this revised semantics; in the second half of the paper we explore one of these in some depth -the problem of abstracting mereotopological relations from noisy video data, such that the sequences of qualitative relations between pairs of objects do not suffer from "jitter". We show how a Hidden Markov Model can exploit this distance based semantics to yield improved interpretation of video data at a qualitative level.
We present a method for unsupervised learning of event classes from videos in which multiple actions might occur simultaneously. It is assumed that all such activities are produced from an underlying set of event class generators. The learning task is then to recover this generative process from visual data. A set of event classes is derived from the most likely decomposition of the tracks into a set of labelled events involving subsets of interacting tracks. Interactions between subsets of tracks are modelled as a relational graph structure that captures qualitative spatio-temporal relationships between these tracks. The posterior probability of candidate solutions favours decompositions in which events of the same class have a similar relational structure, together with other measures of well-formedness. A Markov Chain Monte Carlo (MCMC) procedure is used to efficiently search for the MAP solution. This search moves between possible decompositions of the tracks into sets of unlabelled events and at each move adds a close to optimal labelling (for this decomposition) using spectral clustering. Experiments on real data show that the discovered event classes are often semantically meaningful and correspond well with groundtruth event classes assigned by hand.
Software technology based on reuse is identified as a process of designing software for the reuse purpose. The software reuse is a process in which the existing software is used to build new software. A metric is a quantitative indicator of an attribute of an item/thing. Reusability is the likelihood for a segment of source code that can be used again to add new functionalities with slight or no modification. A lot of research has been projected using reusability in reducing code, domain, requirements, design etc., but very little work is reported using software reuse in medical domain. An attempt is made to bridge the gap in this direction, using the concepts of clustering and classifying the data based on the distance measures. In this paper cardiologic database is considered for study. The developed model will be useful for Doctors or Para-medics to find out the patient’s level in the cardiologic disease, deduce the medicines required in seconds and propose them to the patient. In order to measure the reusability K-means clustering algorithm is used
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.