Waste classifcation

results of prediction
Accuracy
Loss
real and predicted values
results of prediction

Inspired by global warming.

Thus, we built a model, that predicts the category of an object on the photo.

This will dramatically speed up the waste recycling process if implemented in sorting centers. To build the model we read a lot and after we experimentally chose the values, that suited our model.

How does the application work? You can give a photo to our model and it will give you the category of that object. (Now it is just a prototype, so our application does not have a DUI)

How was the application developed? I came up with the idea of classification waste with a model and started searching the information. I found a dataset that suited my idea and started implementing it. Also, I analyzed the codes of other developers, who were working on the same problem as we are.

How to use the application Unfortunately, we do not have a graphical user interface. Our aim was to build a model, which is not meant to work with users but rather with other machines. It is meant to get a photo and predict the category in waste sorting centers to speed up the process of sorting.

Difficulties & Challenges Find the best parameters for the model: number of epochs, dropout rate, learning rate, number of filters and layers, etc.

_Go-to-Market _ You can contact us to know the details. We are open to collaborations

Enjoy

results video

Experimental selection of the values