Development of Intelligent Basket System by Using Image Detection with Convolution Neural Network (CNN)

Conventional shopping methods are often hindered by inefficiencies such as prolonged queues, delayed checkout process and the absence of real-time expenditure monitoring. All these problems contribute to a time-consuming and potentially frustrating shopping experience. Besides, manual item scanning at checkout counters is susceptible to human error leading to inconsistent billing and customer dissatisfaction. All these operational shortcomings raise broader concerns regarding the societal and environmental implications of traditional retail practices. In response to these challenges, this study proposes the development of an Intelligent Shopping Basket (ISB) system designed to streamline the shopping process through automation and real-time data integration. The ISB implements machine learning technique which is Convolutional Neural Networks (CNN) implemented via TensorFlow to facilitate accurate and real-time object detection. The items that are placed into or removed from the basket will be automatically identified, with updates reflected instantly through a mobile application that displays both cart contents and total cost. The system is powered by a Raspberry Pi 4B, which serves as the central processing unit, coordinating hardware components including a webcam, LEDs, a buzzer, and a push button. This embedded solution eliminates the need for conventional point-of-sale systems and external payment devices, thereby offering a modern, efficient, and user-friendly alternative to traditional shopping methods.