iRVision AI Box Locator - Updating the Model

In this video, we will go over how to update the AI Box Locator model to improve its accuracy for your specific application. The default segmentation model is powerful, but real-world conditions like lighting variations, box graphics, tape types, and shipping labels can impact performance. That’s why customizing the model with your own annotated images is key to achieving reliable results.

We’ll start by reviewing the essential requirements to ensure you’re ready to begin. Then, we’ll walk through the process of collecting images—whether live or from logs—and annotating them correctly. You’ll learn how to use the annotation interface to add or remove boxes, adjust corners, and apply best practices for precise labeling. Once your dataset is prepared, we’ll show you how to output and transfer it for training, merge multiple datasets if needed, and register the updated model back into your vision process.

Requirements:

• A PC or laptop meeting minimum specifications (including GPU requirements)
• Licensed AI Box Registration software
• Familiarity with the AI Box Locator vision tool

What you will learn:

• How to capture and select images for training
• Steps to annotate boxes accurately for best results
• Tips for avoiding common mistakes during annotation
• How to register the new model and verify improved accuracy

By the end of this session, you’ll have the knowledge to enhance your AI Box Locator’s performance and reduce errors caused by challenging conditions. This is a practical, hands-on guide designed to help you get the most out of your vision system.