Climate change is becoming a serious concern for our future. With thousands of different brands and products, it becomes difficult to make educated purchases based on the product’s environmental impact. By making the information accessible, we hope to help people make better choices in the items they buy.
What it does
Sustainable Shopper scans an item that a customer is interested in purchasing, uses computer vision to detect what the product is, and outputs information on the environmental impact of the brands that create it.
How I built it
In order to build our webapp, we used HTML, CSS, Pillow, and Flask. To add the image recognition functionality, we used Clarifai’s image recognition API. We put in hundreds of photos to train the image recognition and output the proper description of the product.
Challenges I ran into
Over the course of our project, the biggest problem we needed to solve was figuring out how to send the image uploaded on the HTML frontend to the backend to be processed. Afterwards, we needed to figure out how to send the descriptions to the HTML frontend to update the HTML display. These difficulties were largely in part to the fact that we used Flask, a technology relatively unfamiliar to our team.
Accomplishments that I'm proud of
We are proud of training the image recognition tools and successfully processing and sending the proper output to display on the HTML and learning how to use the Flask framework and Clarifai image recognition all in one weekend together. Also, we are proud of the way we were able to collaborate and efficiently bring our idea to life.
What I learned
Prior to the project, our team had never worked with computer vision. Also, we were mostly unfamiliar with creating web apps with Flask. Creating the project has helped us improve our knowledge significantly.
What's next for Sustainable Choices
In the future, we would like to scale the power of Sustainable Shopper by adding a larger database of photos and categories. Currently, we are limited by Clarifai’s free license to a small number of trainable categories. In order to improve our project, we want to overcome this barrier to make Sustainable Shopper work for a wider variety of products and brands.