Robert Edward Loke

Eddy Loke was born in Woerden, The Netherlands. In 1994, he received the M.Sc. degree in Computer Science Engineering from Delft University of Technology, and, in 1995, the M.A. degree in Psychology with a 1st-degree education ability from the University of Leiden, both in The Netherlands. From 1997 to 2005, he has been working as a researcher at the Vision Laboratory of the University of Algarve in Faro, Portugal, in various EU-funded (Integrated System for Analysis and Characterization of the Seafloor---ISACS---, Automatic Diatom Identification and Classification---ADIAC--- and EXtreme ecosystem studies in the deep OCEan: Technological Developments---EXOCET/D) and Portuguese projects (Parallel Computing; 3D Modeling from Video), in the fields of image and 3D data processing, scientific visualization, computer graphics and pattern recognition. In 1999, 2000 and 2001, he entered the student poster competitions of the OCEANS, MTS/IEEE conferences in Seattle, Providence and Honolulu on his work in the visualization of sparse gridded sonar data sets and, in 2000, he was awarded the 3rd prize. He has been a reviewer for various IAPR conferences, IEEE Computer Graphics and Applications, and Int. J. of Pattern Recognition and Artificial Intelligence.

Professional interests

(3D) image processing and analysis
images in graphics

Application areas

volumetric analysis of sonar seabottom data
diatom identification
range imaging

Professional projects

EC-EXOCET/D (EXtreme ecosystem studies in the deep OCEan: Technological Developments)
3D MODELING FROM VIDEO
EC-MAST3-ADIAC (Automatic Diatom Identification And Classification)
PRAXIS-SETNA-ParComp (Scalable Environments, Tools and Numerical Algorithms in Parallel Computing)
EC-MAST3-ISACS (Integrated System for Analysis and Characterisation of the Seafloor)

Future research projects

Coding:

image and volume coding using adequate sampling and bin- or quadtree-based decoders

Object detection (i.e. segmentation):

shape-based data filtering in an octree
application to multi-feature/textured data volumes (already done for fast segmentation of range images)
quantitative comparison---in terms of CPU time and boundary quality---with other methods, including relaxation labeling

Geometry extraction (i.e. object boundary triangulation). In particular, extensions to the Linking Matched Cubes algorithm:

improve surface patches at object edges by adaptively setting vertex positions of the additional vertices in shortcut configurations
combine 3D discrete edge/corner recognition and boundary detection in order to preserve more object edges/corners
optimize the number of linked matched cubes (i.e. decrease triangle count) by improving the neighborhood decomposition
apply line and volume primitives for modeling of skeletons and to enable object measurements
triangle reduction at large planar surfaces using a multiresolution implementation

Visualization of objects in sonar data sets:

combine with realtime trajectory planning, already on the vessel, in a closed control loop
further optimization
label the boundary continuity at higher tree levels in order to steer triangulations al lower tree levels (multiresolution meshing)

Diatom morphometric analysis and identification:

diatom contour extraction
optimize classifier input
Mahalanobis metric
incorporate other features, e.g. curve invariant signatures or curvature histograms
curve regularization
boost ID rates using other classification strategies (leave-one-out etc.)
optimize parameter and feature selection
application to help define more new diatom species

Range image segmentation:

application to other data sets
curved surfaces
adaptive segmenter tuning
improved scene understanding using interactive visualization of highly optimized (i.e. efficiently coded) surface triangulations
texture mapping and surface shading

PhD-candidate

Progressive volume visualization of incomplete sonar-data sets---From sea-bottom interpolation and segmentation to geometry extraction.

Journal papers

Loke, R.E. and du Buf, J.M.H.
Quadtree-guided 3D interpolation of irregular sonar data sets.
IEEE J. of Oceanic Engineering 29 (2), 457-471, April 2004.
Mann, D.G., McDonald S.M., Bayer M.M., Droop S.J.M., Chepurnov, V.A., Loke, R.E., Ciobanu A. and du Buf, J.M.H.
The Sellaphora pupula species complex (Bacillariophyceae): morphometric analysis, ultrastructure and mating data provide evidence for five new species.
Phycologia 43 (4), 459-482, July 2004.
Loke, R.E. and du Buf, J.M.H.
Diatom classification in ecological applications.
Pattern Recognition 37 (6), 1283-1285, June 2004.

Conference publications and book chapters

S. Nunes, D. Almeida, Loke, R.E. and du Buf, J.M.H.
Polygon optimisation for the modelling of planar range data.
2nd Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA), June 7-9, pp 128-136, 2005, Estoril, Portugal, ISBN 3-540-26153-2.
Loke, R.E. and du Buf, J.M.H.
Small CPU times and fast interactivity in sonar seabottom surveys.
Proc. 2nd Int. Symp. on 3DPVT (3D Data Processing, Visualization, and Transmission), Thessaloniki, Greece, Sept. 6-9, 2004, CDROM.
P.-M. Sarradin, J. Sarrazin, J.F. Cadiou, K. Olu, D. Desbruyères, V. Rigaud, J.F. Drogou, F. Lecornu, J.F. Rolin, R. Vuillemin, M.A. Cambon-Bonavita, E. Sauter, K. von Juterzenka, A. Boetius, R. Serrão Santos, A. Colaço, A. Pascoal, P. Oliveira, B. Shillito, F. Zal, A. Schultz, P. Taylor, D. Lane, R.E. Loke, H. du Buf, C. Waldmann, A. Cormack, L. Sanfilippo, M. Masson. EXtreme ecosystem studies in the deep OCEan: Technological Developments. Proc. 14th Int. Offshore and Polar Engineering Conf., Toulon, France, May 23-28, Vol I, pp 738-746, 2004, ISBN 1-880653-62-1, CDROM.
Loke, R.E. and du Buf, J.M.H.
Interactively visualizing 18-connected object boundaries in huge data volumes.
Lecture Notes in Computer Science 2886, © Springer-Verlag
Proc. 11th Int. Conf. on Discrete Geometry for Computer Imagery, Naples, Italy, Nov. 19-21, 2003, pp. 544-553, ISBN 3-540-20499-7.
Loke, R.E. and du Buf, J.M.H.
Segmentation of range images in a quadtree.
Lecture Notes in Computer Science 2652, © Springer-Verlag
Proc. 1st Iberian Conf. on Pattern Recognition and Image Analysis, IbPRIA 2003, Puerto de Andratx, Mallorca, Spain, June 4-6, 2003, pp. 428-436, ISBN 3-540-40217-9.
Loke, R.E., Bayer, M.M., Mann, D.G. and du Buf, J.M.H.
Diatom recognition by convex and concave contour curvature.

Proc. OCEANS'02 MTS/IEEE Conf., Biloxi, Mississippi, USA, Oct. 28-31, 2002, pp. 2457-2465.
Loke, R.E. and du Buf, J.M.H.
Diatom contour classification by curvature of convex and concave segments.

Proc. 12th Portuguese Conf. on Pattern Recogn., Aveiro, Portugal, June 27-28, 2002, CDROM.
Loke, R.E. and du Buf, J.M.H.
Identification by curvature of convex and concave segments.

Chapter 8 in: Automatic diatom identification; du Buf, H. and Bayer, M.M. (eds), World Scientific, Series in Machine Perception and Artificial Intelligence, Vol. 51, pp. 141-166, Singapore, 2002.
R.E. Loke and J.M.H. du Buf.
Highlighting pipelines offshore Norway: visualization and quantitative analysis.

Proc. OCEANS 2001 MTS/IEEE Conf., Honolulu, Hawaii, USA, Nov. 5-8, 2001, pp. 2688-2695.
R. Lam, R.E. Loke and J.M.H. du Buf.
Smoothing and reduction of triangle meshes.

Proc. 10th Portuguese Computer Graphics Meeting, Lisbon, Portugal, October 1st-3rd, 2001, pp. 97-105 (Portuguese).
R.E. Loke and J.M.H. du Buf.
Sonar object visualization in an octree.

Proc. OCEANS 2000 MTS/IEEE Conf., Providence, Rhode Island, USA, Sept. 11-14, 2000, pp. 2067-2073.
X. Jiang, K. Bowyer, Y. Morioka, S. Hiura, K. Sato. S. Inokuchi, M. Bock, C. Guerra, R.E. Loke and J.M.H. du Buf.
Some further results of experimental comparison of range image segmentation algorithms.

Proc. 15th Int. Conf. on Pattern Recogn., Barcelona, Spain, Sept. 3-7, 2000, Vol. 4, pp. 877-881.
J. Rodrigues, R.E. Loke and J.M.H. du Buf.
Segmentation of 3D data.

Tecno Visao, EST/UALG, No. 10, 2000, pp. 5-8 (Portuguese).
J. Rodrigues, R.E. Loke and J.M.H. du Buf.
Fast segmentation of 3D data using an octree.

Proc. 11th Portuguese Conf. on Pattern Recogn., Porto, Portugal, May 11-12, 2000, pp. 185-189.
R.E. Loke, R. Lam and J.M.H. du Buf.
Fast segmentation of sparse 3D data by interpolating segmented boundaries in an octree.

Proc. 11th Portuguese Conf. on Pattern Recogn., Porto, Portugal, May 11-12, 2000, pp. 431-433.
R.E. Loke, R. Lam, F. Rotaru and J.M.H. du Buf.
Improving 3D visualisation interactivity by using octrees (extended version).

Proc. Workshop on Medical Imaging, Aveiro, Portugal, Oct. 1999, pp. 19-23.

Also published in: Journal of the Dept. of Electronics and Telecomm., Univ. of Aveiro, Vol. 3, No. 2, pp 167-172.

(Presented poster: header, body.)
R.E. Loke and J.M.H. du Buf.
Fast interpolation, segmentation and visualization of 3D sonar seabottom data by using tree structures.

Proc. OCEANS'99 MTS/IEEE Conf., Seattle, Washington, USA, Sept. 1999, pp. 1571-1576.

(Presented poster: header, body.)
T.R. Reed,R.E. Loke and J.M.H. du Buf.
The analysis of underwater acoustic data via 3-D segmentation.

Proc. 33rd Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, USA, Oct. 1999, paper number MA3b-1, pp. 49-52.
R.E. Loke and J.M.H. du Buf.
Volumetric processing of Topas underwater acoustic data.

Proc. OCEANS'98 IEEE/OES Conf., Nice, France, Sept.-Oct. 1998, pp. 742-746.
S.G. Nikolov, R.E. Loke and J.M.H. du Buf.
Interactive 3-D visualization of sonar data using VRML.

Proc. Int. Workshop on Synthetic-Natural Hybrid Coding and Three Dimensional Imaging, Rhodes, Greece, Sept. 1997, pp. 36-39.

Previous publications

R.E. Loke and G. Cembrano.
Neural adaptive control of a bioreactor.

Proc. Intelligent Components and Instruments for Control Applications, SICICA'94 (2nd IFAC Symposium), Budapest, Hungary, June 1994, pp. 182-187.
R.F.T.Brouwer, A.H.C. van der Heijden, F. van der Velde and R.E. Loke.
The statistical relations between position, colour, and orientation.

Chapter 4 in: One feature to rule them all: The dependencies between position, colour, and form; R.F.T. Brouwer, PhD thesis, University of Leiden, pp. 78-109, 1998.