sharktide/recyclebot0

In the world, very little of what is put into your bin actually gets recycled. One of the prime reasons for this is when people put recyclables in the wrong bin, causing the whole recycling truck to get turned away and sent to the landfill. The amount of materials we have left in the environment is decreasing, and what little we put to recycle can also get wasted. This project tries to address that issue by creating artificial intelligence capable of automatically sorting/categorizing recyclables from images. The model that was created as a result was able to categorize pictures of recyclables correctly 80% of the time. It was found that when a picture of a type of recyclable that fits into one of the 5 following categories is taken, (glass, metal, paperboard, plastic-polystyrene-specific, or plastic-other-recyclable), the model was able to classify it correctly 20 out of 25 times, being extremely confident in correct predictions (Avg: 55%). The significance of this work demonstrates how artificial intelligence can start to be a guide for humans to help make the right choices. However, suppose humans start relying on artificial intelligence too much. In that case, it can have possible unintended consequences, such as people regularly overriding their judgment based on what AI ‘thinks’ can cause people to make the same mistakes as AI, which as shown by the results above can be susceptible to making mistakes.1 This project is anticipated to open the doors for lightweight, efficient, easy, and accessible ways to not only decrease the amount of recycling contamination coming from the home or office environment but also to make it easier and faster for users to recycle. Enter recyclebot.

How to use this model in code:

First install tftools (a package made by me):

pip install --upgrade tensorflowtools

Then download the model: (You only have to do this once unless you want to update it)

import tensorflowtools
tensorflowtools.hftools.download_model_from_huggingface("sharktide", "recyclebot0", "sharktide/recyclebot0")

Now, to load the model in code all you have to do is this:

#Loads the model as if you did tf.keras.models.load_model()
import tensorflowtools
model = tensorflowtools.kerastools.load_from_hf_cache("sharktide", "recyclebot0", "tf_model.h5")
model.summary

A Working Prototype can be found at this project's website: https://recyclesmart.verecl.app

Also, the website code is also on github: https://github.com/sharktide/recyclesmart

The model as a .keras archive can be found on the github repository https://github.com/sharktide/recyclebot

Our API can be found at https://huggingface.co/spaces/sharktide/recycleAI-api/tree/main