Santellegrino

• 

  ✨ What a masterpiece ✨

🧠 Inspiration

Covid has shown us how much we need to be prepared for events that may affect us. Indeed, the fact that we were not prepared for this epidemic wave has plunged Europe and the whole world into a total disaster. There was no vaccine available and no one knew what medication to give to people with the virus.

This is why we sincerely believe that the search for new drugs, through clinical trials, deserves a facelift. We are sure that millions of people are willing to contribute to the development of drugs, but there are only a limited number of people who can go to hospitals every day or week to test them, which drastically reduces the number of participants and therefore the completeness and legitimacy of the study. This problem drastically reduces the number of participants and thus the completeness and legitimacy of the study. The marketing of a new drug/treatment can be extremely risky if the data included in the study are not representative of the population using the drug/treatment.

This is why we are introducing a revolutionary method to facilitate the development of clinical trials. From now on, it will be possible to participate in these trials from home.

💻 What it does

Santellegrino is a new approach to online medicine. Thanks to very precise algorithms, Santellegrino retrieves data such as

• Heart rate
• Respiratory rate
• Nostril dilation
• Facial thinning
• Facial enlargement
• Detection of dark circles
• Patient stress which is determined thanks to heart rate & respiratory rate
• Proposes a non-redundant questionnaire to complete the data taken into account by the algorithms

Santellegrino's very foundation is to reassure patients. From now on, they will benefit from daily help and unprecedented accuracy.

Moreover, Santellegrino is extremely practical because it adapts to any type of computer and therefore allows an infinite number of people to participate and test the Santellegrino experience. At a time when health professionals have an extremely heavy workload and are under great pressure, Santellegrino offers to help them.

In short, it is a new way of reassuring the patient and easing the workload of the caregivers.

🔧 How we built it

We made extensive use of python, mainly using :

• The Open CV library
• The Mediapipe library designed by GOOGLE for the detection of facial points,
• Tensorflow,
• Numpy for matrix calculations
  
• ... and a lot of maths !
  

🔥 Challenges we ran into

The major challenge was to decentralise a multitude of actions normally performed by highly skilled people into an algorithm that only takes into account sound and video.

Reading research papers, very advanced documentation because these are areas where few people have worked.

We were also confronted with many mathematical problems such as the change from the time domain to the frequency domain thanks to the Fourrier transform for the detection of blood flow, the creation of a vector marker to be able to analyse the different points of the face, the possibility of comparing the colours according to the matrices we had at our disposal and finally in all the algorithms.

One of the challenges has also been time, as it seems complicated to develop a solution to a major problem in only 24 hours. Some features can therefore be added to the code and will make Santellegrino even more reliable and efficient.

🦾 Accomplishments that we're proud of

• It may seem trivial, but the simple fact that we didn't sleep a single minute in 24 hours makes us proud
• Being interested in a field in which we were new
• Being able to create a usable and meaningful project
• Being able to recognise very accurately
• To have developed a scale and motion invariant system for face size detection
• Be autonomous in the search for information necessary for the project, in particular in the reading of numerous research articles
• To have put into practice new mathematical concepts

✍️ What we learned

First of all, we managed to master Open CV, which is a difficult library to learn because it includes many features.

In addition, we have perfected our ability to search for information on the Internet and in long and tedious documentations. Indeed, the field in which we created our project is extremely difficult to understand.

Finally we managed to create algorithms without using Deep-Learning only with the help of mathematical formulas, a subject we really enjoy :)

✨ What's next for Santellegrino

One of the great strengths of the algorithm and that it is innovative, which we think is very important in a start-up.

Moreover, the fact that this project helps caregivers, patients and in addition the noble cause of medicine, makes it a very interesting idea to develop. Indeed, with a little more time, means and knowledge, we are sure that we could have added a multitude of very interesting data (e.g. patient's temperature which requires a thermal camera).

If we wanted to create a start-up, we could imagine that in addition to the algorithm we could send tools to improve the results, for example a thermal camera (more accurate).

The algorithm can also be very easily extended to an online platform allowing patients to perform their health check-ups and caregivers to retrieve their data in real time.

Moreover, it is possible to envisage that if certain data collected are critical, an ambulance driver is alerted and urgently contacts the person concerned in order to provide assistance if necessary.

Built With

• maths
• mediapipe
• numpy
• opencv
• python
• tensorflow