Floorfy_Computer_vision

Inspiration -

Although we only have basic knowledge of DL, we wanted to expand our knowledge especially in "Computer Vision". This led us to take on the challenge of "floorfy".

What it does -

It classifies the images when passed into the YOLO model (high accuracy when passed a single image and low accuracy when we automate for every tour)

How we built it -

1. We used pre-trained models for the classification --> Done
2. We tried to extract the labels from this model and tried to include them into a DataFrame which already consists of the columns from the "JSON" file so that we could do some feature engineering and train a model. --> In progress

Challenges we ran into -

1. Since most of our team members work on weekends, it was really hard to work on the challenge itself.
2. Our laptops are a bit outdated and were not able to handle the computation.

Accomplishments that we're proud of -

• We only worked for half a day in total because of our jobs and still able to learn and do something rather than just giving up.

What we learned

-We've learned about computer vision, deep learning, some automation.

What's next for Floorfy_Computer_vision

We will try our best to work on weekends to solve the problem and push it to git.

Built With

• googlecolab
• numpy
• opencv
• pandas
• python
• yolo