Shinsegae I&C unveiled "Sparrows ScanFree" on the 28th, which uses artificial intelligence (AI) vision technology to recognize multiple items at once at a self-checkout without scanning each product's barcode—customers only need to place items on the counter.
The checkout solution recognizes multiple items simultaneously with 99.9% accuracy in 0.4 seconds. The company said this cuts customer wait times by up to ninefold compared with existing solutions that recognize products in 1 second with 99.5% accuracy.
The number of recognizable product units has expanded to 5,000 types (SKU). It can recognize a variety of items without barcodes—from general merchandise and fresh foods such as fruit and vegetables to breads of the same product that each have different shapes.
It also includes a function to ask customers again when the AI's judgment is ambiguous. When there are no significant differences in shape or color, making it hard for the AI to decide, it does not answer arbitrarily but presents a set of candidates for the customer to choose from. The AI learns from customer feedback in real time, improving recognition accuracy the more it is used.
"Sparrows ScanFree" has been introduced at a smart store on the SK hynix Icheon campus. Shinsegae I&C is currently discussing rollout with various retail stores, including bakeries, ready-to-eat specialty shops, and household goods stores.
The company said results are also emerging from the rollout of "Sparrows ScanCare," a solution that analyzes self-checkout use in real time with AI vision to guide accurate usage. Introduced by Shinsegae I&C in 2024, "Sparrows ScanCare" detects situations in which a product is not properly scanned due to customer error or barcode issues, prompts the customer to rescan, and, when necessary, sends an alert to an operator.
At one retail store that adopted "Sparrows ScanCare," customer calls for staff dropped to one-fifth of the level seen with typical self-checkouts that compare weights before and after payment, and it was found to have reduced expense losses from missed scans by up to several billion won annually.
Shinsegae I&C's AI vision–based solutions are designed with an Edge AI architecture that processes data such as images and videos in-store without separate servers or external graphics processing units (GPUs). Processing data only on the device is advantageous for information protection, and the solutions can be introduced without changing existing store systems. Introduction expense has been reduced to about one-quarter that of other solutions.
Shinsegae I&C plans to target Asian markets, including Japan, based on results verified in the domestic market.
Chief Executive Yang Yun-ji of Shinsegae I&C said, "With advanced solutions built on extensive experience from unmanned stores, we plan to accelerate expansion into overseas markets as well as the domestic market," adding, "We will evolve into an expandable solution by combining large language model (LLM)-based agent functions to support customer service, product guidance, and store operations."