Industry Challenge

Reading expiration dates on yogurt containers is challenging for traditional vision systems due to several factors. The expiration date is often printed on a curved surface, which causes distorted characters that are difficult for conventional systems to accurately capture and interpret. In addition, varying font sizes and inconsistent contrast, especially with inkjet printing, make reliable detection more difficult. Lighting conditions further complicate the task, as reflections and shadows can obscure the printed information and reduce overall readability.

Application of Deep Learning

To overcome these challenges, senswork implemented a Deep Learning–based solution that adapted by learning from a wide range of yogurt container examples. The system is trained on large datasets of images showing expiration dates, enabling it to accurately detect and localize text, even on curved surfaces with distorted fonts.

Outcomes

Using deep neural networks, the system verifies the legibility and completeness of each character, ensuring that the entire date is clearly visible and interpretable. It is optimized for speed, allowing real-time reading on fast-moving production lines where multiple yogurt cups are processed simultaneously. Unlike traditional systems, it adapts automatically to different backgrounds without requiring manual adjustments.

ROI Calculator

Discover the potential cost savings of robotic automation over a 20-year system life

This calculator compares your current manual labor costs against the total cost of owning and operating a robotic system over its 20-year lifespan.

EXPLORE TODAY

Furthermore, senswork’s Deep Learning model is highly effective at handling low contrast and distortions, making it particularly suitable for inkjet-printed dates where visibility is often reduced and character shapes may vary. Additionally, the solution is scalable and can be adapted to other product types and packaging formats that rely on inkjet printing.

How do you know if you need Deep Learning for your machine vision solutions? Download the Vision AI White Paper for a guideline.