Integrated Circuit Lead Cosmetic Inspection
Deep learning technology helps limit semiconductor defects and improve yield without the use of extensive defect libraries.
Graphical programming environment for deep learning-based industrial image analysis
Machine vision is used throughout the semiconductor manufacturing process to rigorously monitor quality and catch defects. Manufacturers must be vigilant for scratched, twisted, bent, or missing pins. A chip has such low tolerances for error that any flaw, even the most superficial, is cause for rejection. With so many potential defect types, it is inefficient to program an inspection into a rule-based algorithm. Explicitly searching for all defects is too complicated and time consuming. Deep learning algorithms can help limit semiconductor defects and improve yield without the use of extensive defect libraries.
Cognex Deep Learning offers a simple solution to identify all anomalous features, even without training on “bad” images. Instead, an engineer uses the defect detection tool to train the software on a sample of “good” images in unsupervised mode. Cognex Deep Learning learns the normal appearance and position of a chip’s leads and pins and characterizes all features that deviate as defective.