Signature authentfication

We are a team strongly focused on computing. We really enjoy coding diferent algorithms and resource-intensive programms. We recently started reading about neural networks, and decided to do something useful with them. The SignatureAuthentifier is a programm that uses neural networks and a set of signatures to learn how to identify a given signature while not being fooled by forges. While showing a small version of it, a complete programm would be extremely useful in any area where strong security is required. We build it with a combination of Python to code images and C++ to do the computing bulk. The biggest challenge was working with a neural network, since we had never worked with one before. We used a third-party programm as the base, and then built on it to make it specific to solve our situation. Getting the data was sometimes hard, since changes in the light would trigger unintended responses. As a team, we are really proud of what we have done. We have learnt how to work with neural networks from scratch, and been able to make a strongly efficient code, both in time and the size of the data set required. We feel like we really did what we wanted to do in a stylish way. We have learnt how to choose a proper topology, how to spot errors in data sets that make the neuronal network contaminated, and A LOT about how neuronal networks work. The next step would be making an embeded version for things like phones, safes and similars, the gadgets where it would really shine.

Built With

• c++
• python