Home Business Amazon will use AI to detect damaged items

Amazon will use AI to detect damaged items

Amazon will use AI to detect damaged items

Amazon continues to develop advanced technology based on artificial intelligence (AI). In this case for detect damaged items during the packing processso that they are removed from service and do not reach the client.

Previously, this process was the visual verification of the status of the products This was done manually for years. in Amazon warehouses. Up to five people used a six-point visual verification to verify that the products were in good condition for shipment. However, as Amazon explains, it is a process that takes a lot of time for employees because they “Damaged items are rarely found in inventory«.

Tests in Germany

For this reason and with the aim of speeding up this task, a team of scientists was formed Amazon Fulfillment Technologies develops one AI based tool capable of detecting anomalies and flagging defective products before they are shipped to customers. The project started last year when the research team discovered they could use a machine learning model with reference images to teach the system to compare the product it’s looking at with an image of what the product should look like.

To achieve this, scientists apply Computer vision techniques to scan each item that runs through his camp on the outskirts of the German capital. then one A machine learning model analyzes scanss to discover hidden patterns and spot potential damage. If an error goes unnoticed, experts analyze it and teach the AI ​​to recognize it.

three times more effective

«The system is three times more effective than manually identifying damaged products“, said Christopher SchwerdtfegerApplied Science Manager Amazon Fulfillment Technologies, while suggesting that the company plans to transfer this technology to other facilities. «We would like to implement our damage detection software in 12 fulfillment centers in North America and Europe ahead of the Christmas campaign. Once implemented, the technology will help detect defects in more than 40 million products each month.“, he claimed.

The solution is monitored by operations staff who correct the model to make it more accurate. «These models are useful, but sometimes they are fooled. Our employees teach the AI ​​directly how it can make better decisions in the future. It is this collaboration between man and machine that leads to better outcomes for our customers.added Schwerdtfeger.

The Amazon Fulfillment Technologies team not only plans to install the system at other locations Expand system capabilities and go beyond damage detection prior to shipping an order, for example Determining when and where the damage occurred.

No Comments

Leave A Reply

Please enter your comment!
Please enter your name here

Exit mobile version