We help leading manufacturers “see better.” In simple terms, our solution allows them to see unseen, complex problems.
Complex problems require advanced solutions. So, when it comes to seeing and solving complex issues, we believe that our vision solution can deliver better than the rest. While we’re not prepared to give away all of our secrets, here are three unique approaches that allow us to deliver on our #SeeBetter promise.
- Best-in-Class Data Capture
Knowing what pieces of data are required and how and where to collect that data is critical. Many vision solutions capture and share basic images. But they rarely expose the true cause of the problem. Our AI-enabled vision solution captures high-resolution images, frequently using FLIR thermal cameras. Our team can quickly determine where to install the cameras in order to capture the views we need. The images captured are one piece of a larger puzzle. You only have the full picture when all the puzzle pieces come together. That complete view includes critical insights from high-res image data combined with machine and process data, and operator input. This is why we describe our vision solution as advanced. It goes beyond basic image capture. It provides a comprehensive view of a variety of factors that can contribute to variance, drift, and anomalies that erode quality and your bottom line. - Automated Feature Extraction (AFE)
While our ability to capture data stands out, it’s what we’re able to do with it that really sets our solution apart. Feature extraction is an essential process for image data classification. Generic programming can provide basic AFE and image classification but the majority of existing methods only extract low-level features from raw images without any image-related operations. Our patented feature extraction techniques allow you to take a forensic approach to quality inspection and detection. It magnifies and uncovers problems in your process that might otherwise go unseen. With this real-time, comprehensive view, you can react in real-time. - Machine Learning Scale
Let’s say each completed puzzle is one image. Those images are shared with operators and engineers via our HMI and online platform. They confirm what they are seeing by validating or applying labels defined during initial deployment. Once a critical mass of labeled images are collected, our team can train models to detect the issue. Each time we deploy our solution, we’re expanding our model library. This means that duplicating the success across machines or factories comes faster each time. COVID-19 has forced many manufacturers to adjust the working conditions in their facilities. They are moving quickly to adopt innovation as they return to full capacity with fewer boots on the ground. Our solution provides a new level of “remote control” with the ability to monitor, collaborate and integrate with key systems remotely.
We understand that seeing is believing. In our next several blog posts, we’ll provide more context about how these three key approaches make our Eigen solution a stand-out among vision solutions.