Flutter and Machine Learning

In this tutorial I will show a prototype using Flutter and Machine Learning to recognize missing persons.

Let’s separate this tutorial into the following topics:

1- Feeding our Machine Learning
2- Registering missing persons
3- Making the application in Flutter

The whole process may seem complex, but it is not.

1- Feeding our Machine Learning

The first step is to feed our Machine Learning photos of missing people. Each person will have a unique identify, called “label”, with this “label” that we will identify a person and search their previously registered information in the database.

For this I used Firebase AutoML from Google. In AutoML you upload the images and label them, in my example I uploaded more than 20 images for each missing person, the more images you upload the more accurate the result will be.

There are some tips on how to make your Machine Learning more accurate, as this app was for study purposes only (and as I was using the free plan) I didn’t worry much about it.

At the end of your Machine Learning “training” you can try it out. Upload an image of a person and make sure your Machine Learning accurately returned the desired result.

2- Registering missing persons

The second step is to register missing persons information in the Firebase database, such as name, date of birth, details of the disappearance, images, label, etc.

The value of the “label” field must be the same as the value specified in Machine Learning, this is where we will relate the missing person’s information to the Machine Learning result.

3- Making the application in Flutter

The application flow is very simple:

The user uploads the photo of someone who is missing and the application will send the photo to Machine Learning.

Machine Learning will try to find the people (with their labels) that most closely resemble the uploaded picture, in which case I set to return only results with more than 70% certainty.

With labels returned from Machine Learning, I search the database for their information (labels), returning their name, description, photos, etc.

See the final result:

Application source code:

Related posts:

What does darkness mean? It means invisibility. Invisible means it must be invincible. And what is invincible, surely has to be fearful. So, to sum it up: Darkness is Fear. It is still. Remember the…

Turns out December is a good time to go as everyone’s in the mood to party! I’ve had so much fun over the last couple weeks, catching up with and saying cheerio to my old teams in Economic & Social…

Sitting here drinking my black coffee, I’m counting down the hours until I can eat. Well, not really. I’m only 14 hours into my 24-hour fast, but that just means I haven’t eaten since dinner last…