Projects | Michele Morisco

Asteroid Field

Sat, 09 Mar 2024 23:12:39 +0000

Asteroid Field is a little game that resembles the classic arcade game, Asteroids. The entire game is fully 3D and developed using Unreal Engine 5.

Skills acquired:

Basic concepts of Unreal Engine 5 using blueprints.
Implemented UI of the game
Implemented a custom skybox

Unconstrained maximum problem with SGD and BFGS

Tue, 13 Feb 2024 17:37:08 +0000

The Computational mathematics for learning and data analysis exam’s project consists of:

The problem of estimating the matrix norm-2 of A for a (possibly rectangular) matrix, using its definition as an (unconstrained) maximum problem. Using:

A standard gradient descent (steepest descent) approach.
A quasi-Newton method such as BFGS.

Skills acquired:

Using MATLAB
Implementing from scratch optimization algorithm in Python

Team members:

Michele Morisco
Margherita Pensa

Experiments with Augmented Images

Thu, 27 Apr 2023 10:02:49 +0000

I have tested some ways to use the Augmented Images of Google ARCore. Above, you can see the videos about my experiments.

Playing retro games with Dueling DQN

Fri, 24 Mar 2023 12:36:21 +0000

This project was realized for the Intelligent Systems For Pattern Recognition course and the purpose was to implement a Dueling DQN agent to play three games from Atari and NES environment: Pong, Boxing and Super Mario Bros. using OpenAI Gym.

Parallel approach for the Jacobi method

Fri, 10 Feb 2023 09:59:20 +0000

This project was realized for the Parallel and distributed systems: paradigms and models course. The purpose was to provide a parallel implementation of the Jacobi method using both native C++ threads and the FastFlow library.

Mural AR

Sat, 10 Dec 2022 20:14:36 +0000

Mural AR is an experimental AR-based app where the user can create (from an image), edit a digital mural/graffiti, and place it on vertical surfaces like walls and lockers.

In addition, the user can take a photo of his mural and share it on his social media.

This app is developed with Google ARCore technology and OpenCV.

Offensive Language Detection in Social Media

Fri, 11 Nov 2022 19:14:22 +0000

This project was realized for the Human Language Technologies course and the purpose was to develop an offensive language detector that, given a tweet, can recognize whether it contains offensive content or not.

The dataset used to build the models was obtained from SemEval-2019 Task 6.

Cubic Stylization on MeshLab

Thu, 08 Sep 2022 19:13:29 +0000

The 3D Geometry Modeling and Processing exam’s project consists of:

study the cubic stylization proposed in the paper “Cubic stylization” by Hsueh-Ti Derek Liu and Alec Jacobson
implement it via the C++ VCG library as a new filter plugin for the 3D mesh processing software MeshLab

This plugin is in the official MeshLab release

PokéShot

Wed, 06 Jul 2022 22:25:55 +0000

PokéShot is a tech demo where the player can take a picture of some Bug-type Pokémon inside the player’s area to gain points.

This minigame is developed with Google ARCore technology.

CNN on CIFAR-10 with adversarial attack

Sun, 12 Jun 2022 17:18:04 +0000

Implement the own convolutional neural network and training in it on CIFAR-10. When the network is trained, try an adversarial attack on it, using the Fast Gradient Sign method.

Skills acquired:

gain in-depth knowledge of advanced topics of machine learning, pattern recognition, deep learning, reinforcement learning, and understanding their theory and applications.
To be able to individually read, understand and discuss research works in the field.
To implement pattern recognition applications, machine learning, deep learning, and reinforcement learning models using state-of-the-art libraries and tools.
To be able to make limited advancements and contributions to the research in the field.

Assignment of a set of Intelligent System of Pattern Recognition midterms: 3/3

Restricted Boltzmann Machine on MNIST

Sun, 12 Jun 2022 17:07:57 +0000

Implement from scratch an RBM and apply it to MNIST. The RBM should be implemented fully (both CD-1 training and inference steps).

Skills acquired:

gain in-depth knowledge of advanced topics of machine learning, pattern recognition, deep learning, reinforcement learning, and understanding their theory and applications.
To be able to individually read, understand and discuss research works in the field.
To implement pattern recognition applications, machine learning, deep learning, and reinforcement learning models using state-of-the-art libraries and tools.
To be able to make limited advancements and contributions to the research in the field.

Assignment of a set of Intelligent System of Pattern Recognition midterms: 2/3

Convolution of a Laplacian of a Gaussian blob detector

Sun, 12 Jun 2022 16:49:42 +0000

Implement the own convolution of a Laplacian of a Gaussian blob (LoG) detector with an image and apply it to 3-4 images from the dataset.

Skills acquired:

gain in-depth knowledge of advanced topics of machine learning, pattern recognition, deep learning, reinforcement learning, and understanding their theory and applications.
To be able to individually read, understand and discuss research works in the field.
To implement pattern recognition applications, machine learning, deep learning, and reinforcement learning models using state-of-the-art libraries and tools.
To be able to make limited advancements and contributions to the research in the field.

Assignment of a set of Intelligent System of Pattern Recognition midterms: 1/3

Multiomics integration in PANS

Wed, 01 Jun 2022 21:32:28 +0000

Integrating metabolomics and proteomics data from PANS. Performing classification of patients based on the two types of data, and extracting biomarkers that maximize classification performance.

Skills acquired:

Applying computational technologies to industrial size case studies
Learn the main computational technologies that are used by pharmaceutical, nutrition, and biomedical companies to run research and development activities and to design and manufacture products
Introduction to the field
Biomarker identification
Pathway analysis
Modeling and simulation

Team members:

Gennaro Daniele Acciaro
Michele Morisco

Judith

Fri, 18 Feb 2022 14:20:46 +0000

[In development]

A girl opens her eyes realizing she is dead. She will have to obtain enough emotions and fill her empty heart to be judged and finally leave this world…or at least this is what she thinks.

2D platform/puzzle game with Metroidvania features.

Find Judith’s lost memories and gain emotions analyzing each character’s real-time routine. Emotions unlock new memories and abilities, useful for exploring areas and defending against enemies. Judith’s main power is possessing NPCs and objects to achieve a temporary useful ability.

Team members:

Giada Belcastro (2D Artist, Concept Artist & Game/Level Designer)
Michele Morisco (Programmer)

Autonomous Driving System Simulator

Tue, 15 Feb 2022 09:41:31 +0000

Development of a smart application in the field of autonomous driving. In particular, one has been working on bash scripts and add-ons for CARLA Simulator to improve the QoL of E-Team Squadra Corse, a Formula SAE team of the University of Pisa.

Skills acquired:

Software design and development
Software engineering requirements specification, especially Jackson Problem Analysis
Agile methods, particularly Kanban
Collaborative team working
Mob/Ensemble Programming
Reverse Software Engineering
Scripting UNIX/Linux

Team members:

Lisa Lavorati
Michele Morisco
Matteo Ninniri
Andrea Piras

Hisuian Tales

Mon, 14 Feb 2022 22:13:27 +0000

“Hisuian Tales” is a fangame with original gameplay based on Smeargle’s paintbrush tail. The game has 3 levels and a boss battle. Each level contains different puzzles and enemies to defeat using Smeargle’s tail. Signs are designed to recall original Pokèmon moves, use the right one to progress in the levels.

Team members:

Giada Belcastro (2D Artist & Concept Artist)
Michele Morisco (Programmer & Game/Level Designer)

Nook's string - Il Filo di Nook

Mon, 14 Feb 2022 19:56:36 +0000

Like Arianna, Tom Nook has to rely on his precious red string to find the way out! Enter the labyrinth, find the Bells Treasure and run away, get out thanks to the red string with as many Bells as possible. Game developed for an Animal Crossing forum event.

Team members:

Giada Belcastro (2D Artist)
Michele Morisco (Programmer)

Human-Machine Interface Project

Mon, 14 Feb 2022 18:37:07 +0000

Developing both graphic applications and graphic components.

Skills acquired:

Program using the WindowsForms framework as an example of a graphical toolkit
Program graphic components
Program animation systems
Programming interactive graphic systems in HTML5 (JavaScript, CSS and DOM)
Understand the event management system and event scheduling
Manage homogeneous coordinate systems and affine transformations of the view
3D graphics fundamentals

Data Mining Project

Mon, 14 Feb 2022 17:45:52 +0000

The project consists of data analysis based on the use of data mining tools.

Data understanding
Data preparation
Clustering
Classification
Outlier Detection
Time Series Analysis

Team Members

Edoardo Federici
Michele Morisco
Carlo Tosoni