TI’s AM6xA Processors are designed for vision applications requiring intensive image analytics. Retail checkout and scanner applications like item recognition, barcode scanning and decoding, and theft detection benefit from imaging and vision analytics to improve accuracy, speed, and generality to new environments. This application note analyzes a retail checkout demo application, which uses a raw camera sensor and runs a gstreamer-based application with deep learning for object detection, on the AM62A's heterogeneous architecture. The core load across the AM62A processor is used to select a cost-optimized version of the system-on-chip (SoC) and the application's power usage shows SoC consumpion is under 2 Watts active power.