GAP-AI by Synesthesia positions itself within the Industry 4.0 context in optimizing processes in the management of sales floor plans in retail. This is currently a costly and error-prone operation that requires personnel to visually inspect the shelves of the sales floor, check the display situation, product availability, their placement, promotions, and verify contractual agreements with suppliers. The application of artificial intelligence in automating this process has been making significant progress recently. Images of shelves, captured with mobile phones and tablets, are sent to the cloud where they are processed, and product availability and display situation are calculated. One of the main shortcomings of current systems is the processing speed of the acquired data.
The very latest products on the market require several minutes before they can validate the correct acquisition of images and provide the first analysis results. We aim to explore improved solutions by distributing AI-based analysis between the smart terminal and the cloud, reducing response and validation times of the acquired images, allowing the operator a faster execution of the analysis and an immediate remedial reaction if necessary. Specifically, the project’s objective is to define, implement, and test an initial distributed Proof of Concept system for image analysis based on convolutional neural networks (CNN) implemented and optimized on the terminal and in the cloud.
