Procedural Cities with Houdini and Python

Description:

The world of VFX is demanding efforts to improve the efficiency without losing effectiveness in every step of the animation pipeline. The goal of this project is to create an automated tool able to generate 3D city models in a procedural way. The city plan is populated with buildings of various types, using Python scripts to automate the process. It allows artists to paint the skyline and control the height of buildings before the actual generation.

Technical features:

• Python
• Houdini

 

Frame 1 with opposite light

Frame 1 with direct light

Frame 2

The generated city in different light conditions. We can notice the different types of buildings as well as their alignment to give room to eventual roads.

 

This slideshow requires JavaScript.

The correspondence between the rendered city map and the painting map.

 

 

References:

• https://www.pluralsight.com/courses/houdini-python-procedural-cities

Emotion Recognition Using Deep Learning

Description:

This research project is part of the internship I am currently doing at CRNS – University of Plymouth, UK.

I am conducting a research in the field of computer vision using machine learning techniques. In detail, deep learning methodologies are used and applied for the automated emotion recognition of human faces by means of Automatic Facial Expression Recognition (FER). Due to the nature of the research, Convolutional Neural Networks, a special type of Artificial Neural Network, are applied due to their adaptableness in studying image-based inputs.

In the fist part of my internship I worked on a lab-controlled dataset (CK+). Since some weeks, I am dealing with group emotion classification with the Group Affect Database – in the wild. This means that the pictures contain occluded, not-frontal faces, and so on.

 

Repository

 

 

3D Reconstruction and S-Rep Analysis of Retinal Blood Vessels

The pipeline of the system

A screenshot of the 3D viewer based on jReality

Description: I have worked on this project as Bachelor’s thesis work. It focused on biomedical image analysis/processing.
Eye blood vessels features can evidence pathologies such as Hypertensive Retinopathy (HBP) or CADASIL. We built a tool for detecting and monitoring some of these features, in particular vascular alterations in order to understand risk factors and prevent pathology evolutions. The new method allows to quantitatively and qualitatively analyze the specific region of a retinal blood vessel in new fully automated pipeline capable to analyze real vessels extracted from fundus image. In addition, we have developed a 3D viewer useful to simplify the medical consultation.

Team: Luca Surace, Claudio Stamile, Francesco Calimeri.

Technical features:

• jReality (3D graphics engine in Java)
• Strong theoretical background (S-Rep models and CPNS)
• Matlab

 

Waldo – The Asset Quest

Story: Waldo: The Asset Quest is a Third Person Shooter. The player is the sentient robot Waldo who travels around dimensions without any conceptual meaning.

The game represents a great challenge, since there are no health collectables or tutorial. All of this makes the player in a confused state which evolves in the consciousness of his means and game mechanics. In addition, there are many dangers hanging over the player (traps, enemies, bullets, etc.)

Download and Play!

Team: Luca S., Salvatore I., Alessandro D.R. .

Technical features:

• Unity

MagnaPapers (AGILE)

Description: MagnaPapers is a manager for scientific research. Authors can register and create the profile checking their personal articles through a data-integration process from most reliable scientific databases such as ScienceDirect, Scopus and Google Scholar.

Spring and Hibernate have been used for the implementation.

The project was part of “Agile Software Development for Enterprise” exam.

Team: Luca S., Werner K., Alessandro D. R., Salvatore I.

Role: Designer, Programmer.

Repository (private – ask me for an account)

Plant Development Simulator

Description: Java simulator of plant development using Lindenmayer Systems. In particular, we focused on the modelling and simulation of Fraxinus Pennsylvanica using scientific literature references.

The project was part of Modelling and Simulation exam.

Team: Luca S., Werner K.

Role: Designer, Programmer.

Repository

Lava Flow Visualization

Topographic surface with lava visualization. Texture is mixed with brown color

Topographic surface

Another point of view

Description: Lava flows visualization on a real topographic surface. It is an interactive 3D application which displays a topographic surface and a lava flow on it. The application is general, allowing the visualization of datasets of any size. The data is read from files. It is possible to display data from different points of view and navigate in the scene.

The project is part of the Computer Graphics MSc exam.

Role: Designer, Programmer

Team: Luca S.

Technical features:

• C++, OpenGL Core Profile (GL-SL)
• Shader Programming
• Light management: Blinn-Phong model, Gouraud Shading

Repository

Scania Trucks Failures Prediction

Description: The challenge task of this project was to come up with a good prediction model for judging whether or not a vehicle faces imminent failure of the specific component or not. In particular, the dataset consisted of data collected from heavy Scania trucks in everyday usage. The system in focus is the Air Pressure system (APS) which generates pressurised air that are utilized in various functions in a truck, such as braking and gear changes.

The project was part of Data Mining exam.

Project report

Team: Luca S., Alessandro D. R., Daniele P., Temesgen Z.

Robocup Soccer

The Robocup Competition at my University

Our team is playing

 

Description: Robocup Soccer is a tournament that involves many universities around the world. The research goals concern cooperative multi-agent systems in dynamic environments. The aim is to develop an Artificial Team able to play football against another team. We designed and implemented our team in some weeks. It was tested only against the other ones of our university.

Team: Luca S., Alessandro D.R., Salvatore I., Amanuel S.

Role: A.I. Designer, Programmer.

Technical features:

• Java
• Planning Action System (GOAP)

Repository

Pokémon Catch

Me and my friend Pikachu

Description: This is a game for Android devices in which you have to catch the Pokemon around the world. Their position is displayed in the maps and when you are very close to them, they appears in your camera view.

This was a project only for didactic purposes. Then, all the rights belong to their owners.

Team: Luca S., Alessandro D.R., Pierangela B.

Role: Designer, Programmer.

Technical features:

• Android
• Context Sensitive
• Augmented Reality Technology (Metaio)
• GPS