Contact

Projects

SG-t-SNE-Π

Spaceland Embedding of Sparse Stochastic Graphs

SD-DP

Sparse Dual of Density Peaks

Recent & Upcoming Talks

Processing and analysis of scattered data samples, acquired or generated in various applications, typically involve iterative …
Mon, Aug 26, 2019 11:30 AM — 12:00 PM Gottingen, Germany
Poster

To develop a novel robust method for medical image denoising that uses prior anatomical structure information to perform spatially …
Mon, Jul 31, 2017 2:00 PM
Slides

We present LARK, a mobile web application at edu.lark.gr, that provides location-aware information content.
Thu, Jun 9, 2016 1:00 PM
PDF Slides Video Site

Wed, Apr 6, 2016 3:24 PM — 3:44 PM San Jose, CA
Slides

Featured Publications

In IEEE HPEC

We introduce a nonlinear method for directly embedding large, sparse, stochastic graphs into low-dimensional spaces, without requiring vertex features to reside in or be transformed into, a metric space.
PDF Code Project Supplementary
In JOSS

SG-t-SNE-Π is a high-performance software for swift embedding of a large, sparse, stochastic graph/network into a d-dimensional space (d = 1,2,3) on a shared-memory computer, especially on personal laptop and desktop computers.
PDF Code Project Supplementary
In IEEE HPEC

The density peaks (DP) algorithm for cluster analysis, introduced by Rodriguez and Laio in 2014, has proven empirically competitive or superior in multiple aspects to other contemporary clustering algorithms. Yet, it suffers from certain drawbacks and limitations when used for clustering high-dimensional data. We introduce SD-DP, the sparse dual version of DP. While following the DP principle and maintaining its appealing properties, we establish a sparse descriptor of local density as a robust representation.
PDF Project Slides Supplementary
In MedPhy

To develop a novel robust method for medical image denoising that uses prior anatomical structure information to perform spatially variant noise reduction in each anatomical region in the image, without degrading edge and texture information between and within anatomical structures.
Slides
In MedPhy

Automatic segmentation is very valuable for computer aided diagnosis and contouring during radio- therapy planning. When dealing with multiple image sets, such as the different respiratory phases in 4D-CT data, simultaneous co-segmentation of multiple medical images can be very useful for improving the efficiency and ac- curacy of segmentation and structure tracking. This work aims to develop a robust automatic co-segmentation method by adopting and adapting co-segmentation methods developed by the computer vision community for (RGB) video editing, tracking and understanding to medical image processing.
In JOSS

We present LARK, a mobile web application at edu.lark.gr, that provides location-aware information content.
PDF Slides Video Site
In MedPhy

To facilitate adaptive image filtering operations, addressing spatial variations in both noise and signal. Such issues are prevalent in cone-beam projections, where physical effects such as X-ray scattering result in spatially variant noise, violating common assumptions of homogeneous noise and challenging conventional filtering approaches to signal extraction and noise suppression.
Poster Slides

Recent Publications

Quickly discover relevant content by filtering publications.

We introduce a nonlinear method for directly embedding large, sparse, stochastic graphs into low-dimensional spaces, without requiring …
PDF Code Project Supplementary

SG-t-SNE-Π is a high-performance software for swift embedding of a large, sparse, stochastic graph/network into a d-dimensional space …
PDF Code Project Supplementary

The density peaks (DP) algorithm for cluster analysis, introduced by Rodriguez and Laio in 2014, has proven empirically competitive or …
PDF Project Slides Supplementary

To develop a novel robust method for medical image denoising that uses prior anatomical structure information to perform spatially …
Slides

Automatic segmentation is very valuable for computer aided diagnosis and contouring during radio- therapy planning. When dealing with …
See all publications

Diploma Theses

Open

  • Spectrogram segmentation for speaker-independent multi-speaker separation. See https://arxiv.org/abs/1607.02173 and replicate the result in MATLAB first, and then use (our?) clustering algorithms to do spectrogram segmentation for speaker and language independent speaker separation.

  • GPU t-SNE-Pi, όπως αυτό t-SNE-CUDA: GPU-Accelerated t-SNE and its Applications to Modern Data αλλά πολύ πιο γρήγορο χρησιμοποιόντας καλύτερο αλγόριθμο από εδώ Spaceland Embedding of Sparse Stochastic Graphs

  • Non-periodic multidimensional real convolutions with minimal memory use (multi-level Toeplitz product).

  • Intel Vtune profiler: Να μελετηθεί και να εγκατασταθεί η αλυσίδα εργαλείων της Intel για την ανάπτυξη και μελέτη παράλληλων προγραμμάτων, ειδικά η συμπεριφορά τους στη χρήση της ιεραρχίας της μνήμης.

  • Tapir Cilk Compiler: Να μελετηθεί και να εγκατασταθεί o Tapir Cilk Compiler, να μεταγλωττιστούν κώδικες Cilk (όπου χρειάζεται) και να συγκριθεί η απόδοσή τους σε σχέση με τους gcc & icpc compilers σε σημαντικό αριθμό συστημάτων.

  • Implement an RTL-level RISC-5 in Simulink. Demonstrate correctness by running gcc test programs. See MIPS & ARM 8 diploma theses.

  • Julia on a cluster with multiple GPUs. Configure shared and distributed memory programming with multiple GPUs on a cluster using Julia 1.3. Use the cluster via job queue submission or interactive use. Demonstrate the system with simple parallel implementations and record performance on jupyter notebooks.

In progress

  • Real time video stabilization, image stitching, object recognition on NVIDIA jetson nano.

  • Clustering/Community detection MATLAB toolkit with GPU implementations

Completed

Teaching

Aristotle University of Thessaloniki

Electrical and Computer Engineering

  • Graduate Program of Studies

    • Parallel and Distributed Computer Systems and the Internet of Things

  • New Program of Undergraduate Studies

    • ΗΥ3603 Parallel and Distributed Computer Systems
    • ΗΥ3604 Embedded Real Time Systems
    • ΗΥ1501 Operating Systems
    • 011 Computer Organization

  • Old Program of Undergraduate Studies

    • ΗΥ3302 Computer Systems (-2015)
    • ΗΛ0401 Digital Filters (-2019)
    • ΗΥ3402 Programming Techniques (2016)
    • ΗΥ3502 Computer Architecture (2016)