Visual region understanding: unsupervised extraction and abstraction

Gupta, G. 2012. Visual region understanding: unsupervised extraction and abstraction. PhD thesis University of Westminster School of Electronics and Computer Science

TitleVisual region understanding: unsupervised extraction and abstraction
TypePhD thesis
AuthorsGupta, G.
Abstract

The ability to gain a conceptual understanding of the world in uncontrolled environments is the ultimate goal of vision-based computer systems. Technological

societies today are heavily reliant on surveillance and security infrastructure, robotics, medical image analysis, visual data categorisation and search, and smart device user interaction, to name a few. Out of all the complex problems tackled

by computer vision today in context of these technologies, that which lies closest to the original goals of the field is the subarea of unsupervised scene analysis or scene modelling. However, its common use of low level features does not provide

a good balance between generality and discriminative ability, both a result and a symptom of the sensory and semantic gaps existing between low level computer

representations and high level human descriptions.

In this research we explore a general framework that addresses the fundamental

problem of universal unsupervised extraction of semantically meaningful visual

regions and their behaviours. For this purpose we address issues related to

(i) spatial and spatiotemporal segmentation for region extraction, (ii) region shape modelling, and (iii) the online categorisation of visual object classes and the spatiotemporal analysis of their behaviours. Under this framework we propose (a)

a unified region merging method and spatiotemporal region reduction, (b) shape

representation by the optimisation and novel simplication of contour-based growing neural gases, and (c) a foundation for the analysis of visual object motion properties using a shape and appearance based nearest-centroid classification algorithm

and trajectory plots for the obtained region classes.

1

Specifically, we formulate a region merging spatial segmentation mechanism

that combines and adapts features shown previously to be individually useful,

namely parallel region growing, the best merge criterion, a time adaptive threshold, and region reduction techniques. For spatiotemporal region refinement we

consider both scalar intensity differences and vector optical flow. To model the shapes of the visual regions thus obtained, we adapt the growing neural gas for

rapid region contour representation and propose a contour simplication technique. A fast unsupervised nearest-centroid online learning technique next groups observed region instances into classes, for which we are then able to analyse spatial

presence and spatiotemporal trajectories. The analysis results show semantic correlations to real world object behaviour. Performance evaluation of all steps across

standard metrics and datasets validate their performance.

YearAug 2012
FileGaurav_GUPTA.pdf
Publication dates
CompletedAug 2012

Related outputs

Image quality optimization, via application of contextual contrast sensitivity and discrimination functions
Fry, E., Triantaphillidou, S., Jarvis, J. and Gupta, G. 2015. Image quality optimization, via application of contextual contrast sensitivity and discrimination functions. SPIE Electronic Imaging: Image Quality and System Performance XII. San Fransisco Jan 2015 SPIE.

Contrast sensitivity and discrimination in pictorial images
Triantaphillidou, S., Jarvis, J. and Gupta, G. 2014. Contrast sensitivity and discrimination in pictorial images. in: SPIE Proceedings 9016, Image Quality and System Performance XI SPIE.

Defining human contrast sensitivity and discrimination from complex imagery
Triantaphillidou, S., Jarvis, J., Gupta, G. and Rana, H. 2013. Defining human contrast sensitivity and discrimination from complex imagery. in: SPIE proceedings: Optics and Photonics for Counterterrorism, Crime Fighting and Defence IX; and Optical Materials and Biomaterials in Security and Defence Systems Technology X, 89010C SPIE.

Contrast sensitivity and discrimination of complex scenes
Triantaphillidou, S., Jarvis, J. and Gupta, G. 2013. Contrast sensitivity and discrimination of complex scenes. in: Burns, P.D. and Triantaphillidou, S. (ed.) Image Quality and System Performance X SPIE.

Evaluation of perceived image sharpness with changes in the displayed image size
Park, J.Y., Triantaphillidou, S., Jacobson, R.E. and Gupta, G. 2012. Evaluation of perceived image sharpness with changes in the displayed image size. in: SPIE proceedings: Electronic Imaging: Image Quality & System Performance IX, 82930J SPIE.

Visual scene busyness measures through a region growing spatial segmentation
Gupta, G., Psarrou, A., Triantaphillidou, S. and Park, J.Y. 2012. Visual scene busyness measures through a region growing spatial segmentation. in: Lewis, C. and Burgess, D. (ed.) Optics and Photonics for Counterterrorism, Crime Fighting, and Defence VIII SPIE.

Region analysis through close contour transformation using growing neural gas
Gupta, G., Psarrou, A., Angelopoulou, A. and Garcia Rodriguez, J. 2012. Region analysis through close contour transformation using growing neural gas. in: WCCI 2012 IEEE World Congress on Computational Intelligence, IJCNN, Brisbane, Australia, 10-15 June 2012 IEEE . pp. 1-8

Image segmentation based on semi-greedy region merging
Gupta, G., Psarrou, A. and Angelopoulou, A. 2012. Image segmentation based on semi-greedy region merging. in: IET Conference on Image Processing (IPR 2012), University of Westminster, UK, 3-4 July 2012 Stevenage Institution of Engineering and Technology (IET). pp. 211-214

Hand gesture modelling and tracking using a self-organising network
Angelopoulou, A., Garcia-Rodriguez, J., Psarrou, A. and Gupta, G. 2010. Hand gesture modelling and tracking using a self-organising network. The 2010 International Joint Conference on Neural Networks (IJCNN) . Barcelona 18 Jul 2010 IEEE .

Towards symbolic reasoning from subsymbolic sensory information
Bolotov, A., Gupta, G. and Psarrou, A. 2010. Towards symbolic reasoning from subsymbolic sensory information. in: Bolotov, A. (ed.) Proceedings of the automated reasoning workshop 2010: bridging the gap between theory and practice. ARW 2010 University of Westminster.

Tracking gestures using a probabilistic self-organising network
Angelopoulou, A., Psarrou, A., Garcia Rodriguez, J. and Gupta, G. 2010. Tracking gestures using a probabilistic self-organising network. in: 2010 International Joint Conference on Neural Networks (IJCNN), Barcelona, Spain, 18-23 July 2010 IEEE . pp. 1-7

Generic colour image segmentation via multi-stage region merging
Gupta, G., Psarrou, A. and Angelopoulou, A. 2009. Generic colour image segmentation via multi-stage region merging. in: Proceedings of the 10th Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS '09) IEEE . pp. 185-188

Active-GNG: model acquisition and tracking in cluttered backgrounds
Angelopoulou, A., Psarrou, A., Garcia Rodriguez, J. and Gupta, G. 2008. Active-GNG: model acquisition and tracking in cluttered backgrounds. in: VNBA '08: Proceeding of the 1st ACM workshop on vision networks for behavior analysis ACM. pp. 17-22

Robust modelling and tracking of NonRigid objects using Active-GNG
Angelopoulou, A., Psarrou, A., Gupta, G. and Garcia Rodriguez, J. 2007. Robust modelling and tracking of NonRigid objects using Active-GNG. in: IEEE Workshop on Non-rigid Registration and Tracking through Learning, NRTL 2007, in conjunction with ICCV 2007, 14-21 October 2007, Rio de Janeiro Los Alamitos, USA IEEE . pp. 1-7

Nonparametric modelling and tracking with Active-GNG
Angelopoulou, A., Psarrou, A., Gupta, G. and Garcia Rodriguez, J. 2007. Nonparametric modelling and tracking with Active-GNG. in: Lew, M., Sebe, N., Huang, T.S. and Bakker, E.M. (ed.) Human-Computer Interaction: IEEE international workshop, HCI 2007, Rio de Janeiro, Brazil, October 20, 2007; proceedings Berlin Springer.

Permalink - https://westminsterresearch.westminster.ac.uk/item/8z326/visual-region-understanding-unsupervised-extraction-and-abstraction


Share this
Tweet
Email