ShopEz

Live counter for crowd analysis
Mask detection

These are troublesome times now and all small businesses have taken a drastic hit due to the fall in economy. This is a little something that we wanted to do to help them out

Using our website customers can add things to the cart and confirm the order so that the shopkeepers can get it ready before they come for pickup, and they will get a text message with the order number which should be shown to collect the goods. And also customers can see the live count of the number of people in the shop to decide when to go shopping incase they want to do it in person. For the shopkeeper's part, mask detection and social distancing maintenance will be monitored by an AI and will alert the shopkeeper if someone fails to practise wearing masks or social distancing. We have aslo added an automated thermostat, illumination control and sanitizer dispenser so it saves time and money for the shopkeeper.

User interface was built predominantly using python and eel was used for the website. AR was implemented using Echo AR. The video was recorded using imutils and all models for detection were trained using Tensorflow's SSD MobileNet V2 with GPU support to reduce training time. All hardware simulations were performed in TinkerCAD. For live count of people in the shop, we used google python api to send it to the server.

We had to learn to use Echo AR as it proved to be very useful and easy to implement. Training models for face mask recognition and people recognition was a pain XD, even with GPU support.There were cases of overfitting with the people detection model, which is due to training for more steps than necessary. Overall combining all of the individual contributions from my teammates was difficult but fun.

We had to observe a lot on how consumers and amall businesses are affected during covid and had to think of a solution and it works!! We couldn't be any happier that it worked. And also we pulled it off within two days without sleep!!!

Learning AR was the most fun and interesting part. We learnt how to implement multiple projects into one, as ours is both hardware and software. And prime of all, teamwork.

Shop Ez will continue to strive to improve itself by implementing an easier UI and many more added features for the business owner and the customer.