This web application aids in mobile inventory system management. The frontend of the mobile device shows a stream of things being displayed; a deep learning model uses this data to identify the objects and automatically update the inventory list's frontend display with the most recent inventory count.
● Using state of the art Yolov5 model with detection speed as low as 5ms per frame.
● Built the frontend using Streamlit.
● Using SQLite3 as a Database for managing inventory. .
Utilizing streamlit_webrtc, which facilitates real-time communication, we will be able to detect images in real time by executing detections on the frames captured from the real-time video in the backend. You could also use drag-and-drop or image upload to carry out detections on images, as seen below.Using three fields — Date, Beverage Name, and Count — we are storing results on our sqlite database, which aids the end user in maintaining inventory and verifying quantities day to day. as shown below
If you would like to experiment with the custom dataset yourself, you can download the annotated data on Kaggle and the code at Github the web app is deployed Here .