[1] Dimitris Kastaniotis, Ilias Theodorakopoulos, Spiros Fotopoulos, "Pose-based gait recognition with local gradient descriptors and hierarchically aggregated residuals," J. Electron. Imaging 25(6), 063019 (2016), doi: 10.1117/1.JEI.25.6.063019

[2] Dimitris Kastaniotis,Ilias Theodorakopoulos, George Economou and Spiros Foropoulos Gait based recognition via fusing information from Euclidean and Riemannian manifolds, Pattern Recognition Letters, Volume 84, 1 December 2016, Pages 245-251, ISSN 0167-8655,

[3] Dimitris Kastaniotis, Foteini fotopoulou, Ilias Theodorakopoulos, George Economou, Spiros Fotopoulos, "HEp-2 cell classification with Vector of Hierarchically Aggregated Residuals",accepted to Pattern Recognition
Code is available: here

[4] Dimitris Kastaniotis, Ilias Theodorakopoulos, Christos Theoharatos, George Economou, Spiros Fotopoulos, A framework for gait-based recognition using Kinect, Pattern Recognition Letters, Volume 68, Part 2, 15 December 2015, Pages 327-335, ISSN 0167-8655,

[5] "HEp-2 Cell classification using descriptors fused into the dissimilarity space", Ilias Theodorakopoulos, Dimitrios Kastaniotis, George Economou and Spiros Fotopoulos has been accepted in for publication in International Journal on Artificial Intelligence Tools.

[6] "KAD - An Intelligent System for Categorizing and Assessing the State of Patients with Multiple Sclerosis", Spiros Fotopoulos and Dimitrios Kastaniotis, ERCIM issue 95, October 2013, pp.25, Special Theme "Image Understanding " available online

[7] Ilias Theodorakopoulos, Dimitris Kastaniotis, George Economou, Spiros Fotopoulos, HEp-2 cells classification via sparse representation of textural features fused into dissimilarity space, Pattern Recognition, Volume 47, Issue 7, July 2014, Pages 2367-2378, ISSN 0031-3203. doi:

[8] Ilias Theodorakopoulos, Dimitris Kastaniotis, George Economou, Spiros Fotopoulos, Pose-based human action recognition via sparse representation in dissimilarity space, Journal of Visual Communication and Image Representation, Volume 25, Issue 1, January 2014, Pages 12-23, ISSN 1047-3203, Visit Project Page


[1] Dimitrios Kastaniotis, Gerasimos Kartsakalos, Spiros Fotopoulos, Panagiotis Papathanasopoulos, "Using Kinect for Assesing the state of Multiple Sclerosis patients," To appear in Mobihealth 2014.

[2]Ilias Theodorakopoulos, Dimitris Kastaniotis, George Economou and Spiros Fotopoulos "HEp-2 cells classification using morphological features and a bundle of local gradient descriptors ", To appear in the First workshop on Pattern Recognition Techniques for Indirect Immunofluorescence Images. August 2014, Stockholm Waterfront, Sweden.

[3] Dimitrios Kastaniotis, Ilias Theodorakopoulos, George Economou, Spiros Fotopoulos, "HEp-2 cells classification using locally aggregated features mapped in the dissimilarity space," bibe, pp.1-4, 13th IEEE International Conference on BioInformatics and BioEngineering, 2013 Available online in pdf.

[4] D. Kastaniotis, I. Theodorakopoulos, G. Economou and S. Fotopoulos, Gait-based Gender Recognition using Pose Information for Real Time Applications", Digital Signal Processing (DSP), 2013 18th International Conference on , vol., no., pp.1,6, 1-3 July 2013

[5] F. Fotopoulou, D. Kastaniotis, I. Theodorakopoulos and G. Economou, "Shape Representation Via the Generalized Geodesic Median Point", to appear in ICPRAM 2013

[6] I. Theodorakopoulos, D. Kastaniotis, G. Economou and S. Fotopoulos. "HEp-2 Cells Classification via fusion of morphological and textural features". In proc. of IEEE 12th International Conference on BioInformatics and BioEngineering (BIBE), November 2012 Visit Project Page

** As Undegraduate Student

[1] "A System for Acquiring, Transmitting and Distributed EEG Data Processing." D. Kastaniotis, G. Marangos, N. Fragoulis and A. Ifantis, MEDICON 2010, IFMBE Proceedings 29, pp. 843- 846, 2010 **

You can also view my Google Scholar Profile here

Master Thesis: Title: "Shape Retrieval Using Diffusion Processes"


In this Thesis Diffusion Maps were used in order to extract bending invariant descriptors to match shapes from the MPEG-7 planar shape database. The efficiency of Diffusion Distances into exploring the Geometry of any data set modeled as graph as well as the superiority of Diffusion over geodesic distances is studied here. In this context, every shape is represented as a graph and from this graph a Markov Matrix is formed. A combination of Spectral Graph Properties among with a Markov Process provides as a multiscale analysis. The Eigenvectors of the Markov Matrix are used to re- represent the data in every time step of the Markov Matrix Iteration. From this mapping the pairwise spectral distances are used to form a histogram descriptor which is later used in order to match shapes using the L1 distance. For first time here the particular properties of small, medium and large scales are studied. A sigmoid function is used in order to exploit the information contained in medium and large scales by weighting the histograms that correspond to these scales with larger values than those histograms derived from small scales. The retrieval score in both Kimia 99 and MPEG-7 (63.75%) databases overpasses geodesic histogram descriptors.
Keywords: Diffusion Maps, Diffusion Distance, Shape Retrieval, MPEG-7

You can read my Thesis (in Greek) here

Graduate Program "Electronics and Information Processing"

Supervisor: Professor Spiros Fotopoulos