:::: MENU ::::

Global illumination

Voxel Cone Tracing


On the November 24th, Dragan Okanović talked about global illumination and real-time techniques that exist and that are used in game production.

Some of the covered techniques were: pathtracing, radiosity, irradiance volumes, [C]LPV, RTV, SVOGI, RRF. Presentation slides with additional notes (in English) can be found at slideshare or GDrive.

Procedural generation of the buildings

On the November 12th, Nemanja Lučić talked about procedural generation of the settlements using formal grammar and gave the demonstration using the sample app that he wrote as a part of his final graduate project.

Talk covered brief overview of the formal grammar and how it’s used for procedural generation. The rest of the talk was focused more on the exact implementation of the system and various challenges that were discovered during the work on the project and ways to solve them. Finally, demo app was presented (code to be uploaded soon). Presentation slides (in Serbian) can be found on GDrive. Project’s code can be found in this bitbucket repo.

Visit from BMIT

On Tuesday, May 12th, members of BMIT research laboratory came to visit to the LARGO and they demoed projects that they are working on. They presentad brief introduction of EMG usage and demonstrated how the data is measured and filtered. The talk also covered practical use of those signals, how they’re used to predict deviation of normal behavior and how to use signals to send meaningful commands to the computer. Backward approach, where analysis of such signals is used to stimulate muscles was also presented.

Need for application of biomedical technology is real and present, and so we hope that more people will get involved in this area of IT. Practical use of BM IT is diverse and life changing, and it could also provide human beings with some amazing abilities, that could enhance our lifestyle.

We would like to thank Nadica Miljkovic, Matija Strbac, Vladimir Kojic, Milica Isakovic and Sergej Kasavica for their time and inspiration.

HFPaC talk

On the March 17th, professor Djordje Djurdjevic talked about HFPaC (height-field parallel compression) algorithm.

Talk covered origin of the problem, previous methods to solve problem and presented a new method for compressing height-maps using a Bezier surface approximation for patches. Presentation slides (in Serbian) can be found: here




At the GDC (Game Developers Conference) 2015 few new technologies were presented:

Vulkan API:
Vulkan, sometimes referred to as glNext, is a low-level cross-platform 2D and 3D graphics API first announced at GDC 2015 by the Khronos Group. The foundation on which Vulkan was built is derived from components of AMD’s Mantle.



Unreal Engine



On the February 25th 2015, Jovan Radivojsa talked about the Unreal Engine.

Presentation covered history of the Unreal Engine development, basic concepts and hierarchy of the objects inside the engine as well as the UI of the editor. Tutorial of how to create a simple project with an engine was demonstrated through a series of short videos.

Files from the presentation (in Serbian) can be found here: [link]

YouTube videos that were shown during the presentation that cover introduction to the Unreal Engine scripting, UI and editor can be found here.

Pathtracing and Global Illumination



On the January 28th 2015, Dragan Okanovic talked about path tracing and global illumination.

Presentation briefly covered problems that computer graphics is trying to solve, such as what light actually is and how to model it, and rendering equation that is trying to solve (Kajiya). Whitted raytracer implementation, Monte Carlo approximation for GI were mentioned and also various ways to more efficiently get to the solution with MC methods – Metropolis Light Transport, Bidirectional pathtracing, Photon Mapping and various acceleration structures (BVH, K-d tree, Octree) were also discussed.

Files from the presentation (in Serbian, with notes) can be found here: [link]

Intro to Neural Networks and Genetic Algorithms

Neural Network example

On the 24th December , Aleksa Pavlovic presented an intro to neural networks theory and genetic algorithms. He talked about theory and usage of NNs and GAs in the industry, and he explained how he implemented AI in his course assignment that trained itself over time.

Files from the presentation (in serbian) and the full game with implemented GA-trained AI can be found here: link