DetectIT

• 

  Object recognition in cities

• 

  Researched different computer vision tasks

• 

  Face recognition at work

Inspiration

Being curious about how face recognition tools work, I wanted to learn about its implementation and dig in more into machine learning, thus I decided to start with object recognition and labeling.

What it does

Recognize and label objects such as soda cans and water bottles in real time, as well as their contour with surrounding boxes on the front-end.

How I built it

I first created my own data by equally taking 80 pictures of different objects, and using IBM's Watson ML cloud platform, I was able to efficiently train my data. The React app helped in displaying the model in action using my laptop's camera.

Challenges I ran into

As it was my first time using Watson Machine Learning, I had a hard time training my data to have high accuracies, which took the most time.

Accomplishments that I'm proud of

I am most proud of tackling a field of interest, and also of completing a project with new technologies and tools.

What I learned

I learned how to train data on IBM's Watson ML in order to recognize and label objects, as well as React (including the Effect hook).

What's next for DetectIT

As I tackled my curiosity in machine learning with a smaller-scale project with object recognition, I am planning on expanding the project, to do face recognitions and include a lot more diverse training data.

Built With

• ibm-watson
• javascript
• react