This project was created using a Raspberry Pi model B+, python, and a bunch of other cool stuff. It started as a fun thing to do and we ended up winning an innovation contest. The aim of this project is to design a solution to get important data from floors.
Speech Therapy apps aren't very common, speech therapy apps written in Portuguese are even less. This application is being constructed in partnership with a speech therapist and it contains useful exercices for people in need.
Creating urban environments in games, films and other virtual environments is a complex task that can take a lot of time, we just need to think about all the elements we can find in cities (terrain, roads, buildings, trees, etc...) to see how complex it is. My BSc final project goal is to create a new procedural algorithm to create random cities(based in some basic inputs, like desired city size), and to build a real time engine to show the cities generated by the algorithm. This is currently being developed in C and OpenGL.
Like the ALBA project I really enjoyed working on this one. This was a team project to program an handwritten digit recognizer. With the access to a data base with 3500 digits we started by doing some previous treatment to the images. After that we extracted 32 features from each image and we made a normalized database with them. We have implemented 2 classifiers, the K-NN and Naive Bayes. After that we have implemented the Genetic Algorithm to reduce the features needed to classify a new image and to improve the successful classification rate. After all this, with K-NN, we can successful classify 93% of the new handwritten digits with less than 20 features to analyze.
SecuriText was a team project for my Computer Science project. We made a secure text editor that can encrypt and decrypt any file with RC4 (Rivest Cipher 4), 3DES (Triple Data Encryption Standart) and AES (Advanced Encryption Standart). It can also sign files and create certificates. The application was programmed in Python and the interface was design using GTK.
This is a command line application written for my Computer Security subject. It can encrypt and decrypt any file with DES (Data Encryption Standart) using CBC (Cipher Block Chaining) mode. I was allowed to use the OpenSSL library to implement the encryption with DES in ECB (Electronic Code Book) mode, the CBC was implemented by hand using the output from ECB. The PBKDF (Public Based Key Derivation Function) was also implemented by hand.
The Rich Text Format Editor showned in picture was my Human-Computer Interaction project. The main objective of the project was to built a good looking and functional text editor, combining the right colors and icons with the application layout. This application was programed in Java, using Java Swing.
This was a team project to enter in the contest "Microsoft Pizza Night 2013 - Spring Edition", we won the second place.
My role in the project was game design and programming. It is already published in Chrome Web Store. Link here
This was our Formal Languages and Compilers subject final team project.
We built an high level language to run in MIPS. We have defined the grammar and programmed the: Lexer(using FLEX), the parser(using Bison), the Abstact Sintax Tree, the error handling system and the code generation for assembly.
At the end we had it generation the code to assembly to declare intereger and boolean variables, print to stdout, read from stdin, atributions and arithmetic operations.
We also made a smart htmlizer to ident ALBA code(picture related). This was for sure one of my favourite projects I have made.
One of the branches of forensic science is the recognition of individuals for purposes of criminal investigation, based on a set of information about physical characteristics. The aim of this work was to design and implement a system of research subjects based on an huge characteristics data base. To search for the correct crime suspect in usefull time I've implemented a valuated tree based on their characteristcs. After that I've implemented the k-nn algorithm to get the k suspects with the characteristics similar to a standard input user and other algorithm to show all the suspects which do not differ more than a certain input distance from the pattern introduced.