Mango ripeness testing and product suggestions

Inspiration

The staggering fact that over 1/3 of all food produced globally goes to waste drove our initiative. In the UK alone, 9.5 million tons of food waste are discarded annually, amounting to 391 grams per person daily. Recognizing the complexity of assessing mango ripeness, we saw an opportunity to reduce fruit wastage.

What it does

Our Mango Ripeness Testing and Product Suggestions tool uses AI to accurately determine the ripeness of mangoes and suggests products or recipes best suited for their current state.

How we built it

Using a blend of computer vision powered by CNNs and TensorFlow, paired with the z16 Telum processor and IBM Z Deep Learning Compiler, we trained our model on various stages of mango ripeness. The system was further enhanced with a database of recipes and culinary suggestions.

Challenges we ran into

Differentiating between subtle ripeness stages was tough. Also, ensuring the system worked under different lighting conditions and on various mango varieties posed challenges.

What we learned

The importance of a diverse dataset, especially when dealing with natural products. We also realized the potential of integrating AI into everyday decision-making processes.

What's next

Expansion to other fruits and vegetables, enhancing the product suggestion feature, and a potential mobile app for on-the-go ripeness checks!

Built With

• apriori
• python
• tensorflow
• yolo8