GSoC Projects 2014

We participated as a mentoring organization in 2014, for our 4th consecutive year, with 13 projects proposed and mentored by our Nodes and international community.

Data-centric provenance capture with Sumatra
Implementing a data-centric approach to provenance capture in Sumatra, usability and interoperability of the reproducible research tool are enhanced. Establishing infrastructure for data-process associations in the first part of this project will enable the addition of data-centric views to the software's interfaces in the second part, providing a new and important mode of sharing, reproducing and interacting with computational results with Sumatra.

Student: Felix Hoffmann (Germany)
Mentor: Andrew Davison (France)

Dropbox upload client for ImageJ
ImageJ is an open source Java based image processing program extensively used in life sciences. Our project aims at developing an ImageJ plugin which will provide users with a Graphical user interface (GUI) to upload and download datasets (images, docs, folders, codes etc.) from their Dropbox accounts (using Dropbox Core APIs for Java).

Student: Atin Mathur (India)
Mentor: Dimiter Prodanov, Belgian INCF node

Enhancing the Android Conference App
The project focuses on completing and enhancing the G-Node conference application for Android Platform, which currently provides basic functionality and electronic conference guide for conference participants. The enhancement would focus on extending it's functionality such that it acts as an all-in-one tool for conference participants and delegates, by adding features like option for taking notes against some abstract, synchronization of data with server, itinerary planner, and much more.

Student: Shumail Mohy-ud-Din (Pakistan)
Mentors: Christian Garbers & Christian Kellner (Germany)

Mobile application for Electrophysiology Databases
Implement a framework, via extending the current one implemented for relational electrophysiology databases, to generate a mobile layout according to data structure defined in a NoSQL electrophysiology database.

Student: Rahul Kadyan (India)
Mentor: Petr Ježek (Czech Republic)

NeuroStars: A question and answer site for the neuroinformatics community
Improve Neurostars.org website in order to benefit the Neuroinformatics community. Make any new feature compatible with the original Biostar.org codebase so the 2 repositories could be easily merged and also the Bioinformatics community can benefit.

Student: Paolo Coffetti (Italy, Netherlands)
Mentors: Roman Valls Guimera (INCF Secretariat) & Satrajit Ghosh (US)

Object Oriented Python Kernel Current Source Density toolbox
Kernel Current Source Density Method is a recently developed method for estimating the density of trans-membrane current sources, which is useful for the study of neuronal synaptic dynamics. The method involves inverting a forward modeling scheme which is different depending on the dimensionality of electrodes distribution. The aim of this project is to create an object oriented python implementation of the method with a coherent design, that would also address the performance issues.

Student: Grzegorz Parka (Poland)
Mentor: Daniel Wojcik (Poland)

Open source, cross simulator, large scale cortical models
In this project we attempt to further the efforts of the neuroinformatics community in the direction of having a library of open neuronal models for the benefit of science.

Student: Ramón Martínez Mayorquin (France)
Mentors: Andrew Davison (France) & Padraig Gleeson (UK)

OpenWorm unified data model
OpenWorm relies on myriad storage formats to accomplish its goal of simulating C. Elegans. Although components of the project are served well by these formats, they must be combined into cohesive presentations. Currently, combining distinct data sources is ad hoc and susceptible to breakage when new storage formats are adopted. I propose to virtualize access to the different storage formats into RDF to improve this situation.

Student: Mark Watts (US)
Mentor: Stephen Larson (OpenWorm)

OpenWorm: Replaying of neuroscience simulations in Geppetto
Geppetto is the simulation engine developed by OpenWorm, an organization attempting to build the first cell-by-cell simulation of the nematode C. elegans. The project aims at enabling Geppetto to replay simulations from popular neural simulators (Neuron, Brian, ...). Firstly, the output of these simulators will be investigated with simple neuron models. Secondly, Python scripts will be developed that transform the output to Geppetto's format. Finally, documentation for researchers will be added.

Student: Johannes Rieke (Germany) Mentor: Matteo Cantarelli (OpenWorm)

Optimization of parallel computation in MOOSE
The Multiscale Object-Oriented Simulation Environment (MOOSE) has recently implemented a new MPI-based parallelization scheme that supports automatic setup-time load balancing and transparent object access across nodes. This project aims to improve parallelization in MOOSE using CUDA in order for better performance on GPGPUs.

Student: Vivek Vidyasagaran (India)
Mentors: Upinder S. Bhalla (India)

Real-time vectorization of brain atlases
The overarching goal of this project is to extend INCF Scalable Brain Atlas with a service in which a user can select an online atlas, upload a brain scan, and see the regions of the atlas projected on the scan. Two main ingredients for achieving this are (1) Image registration and, (2) Bitmap to vector tracing. Many good implementations for image registration exist, but for vector tracing the main candidates (AutoTrace, PoTrace) have severe shortcomings stated in the proposal with the solution.

Student: Dhruv Kohli (India)
Mentors: Rembrandt Bakker (Netherlands) & Piotr Majka (Poland)

Volumetric Time series visualizer for TVB and updating other visualization
The main goal of this proposal is to implement a time-series visualizer for 4D data in TVB, which would enable users to work with their functional MRI data directly from the browser. Then I will work on rewriting some of the visualizers that are currently implemented using MatplotLib and MPLH5, backed with web-friendly tools.

Student: Robert Parcus (Italy)
Mentor: Lia Domide (The Virtual Brain)