SAN FRANCISCO--(BUSINESS WIRE)--Today, February 4th, 2020, Nanotronics, the inventor of a platform that combines AI, automation, and sophisticated imaging for industrial inspection, will present two papers at SPIE, the International Society for Optics and Photonics, for Photonics West, held in the Moscone Center.
“The two works presented at SPIE West represent Nanotronics’ continual drive to combine Artificial Intelligence and Optical measurements,” says Vadim Pinskiy, Vice President of Research and Software Development at Nanotronics. “We show the ability to construct more accurate three-dimensional surface profiles and advanced automatic critical dimension measurements.”
As the photonics field continues to advance, so do Nanotronics’ industry agnostic system of nSpec® tools.
“Advancing inspection capabilities in a single tool is vital for any process that is undergoing radical changes, such as in the photonics industry, where we have focused our research efforts,” says Dr. Matthew Putman, cofounder and CEO of Nanotronics. “These additions to our capabilities further reduce variation inherent in using multiple metrology tools, while also exploring features that have never been analyzed before.”
Advanced Approaches to Critical Dimension Measurement in High-Throughput Industrial Processing
Time: 11:15 AM - 11:30 AM
Authors: John Cruickshank, Jeffrey Witz, Toni Ivanov, Vadim Pinskiy, Matthew Putman
Critical Dimensions Measurement is a principle Quality Control method for industrial semiconductor production. Though an array of instruments is available for characterization, these instruments are often specialized for a narrow band of samples. They are also expensive and lack throughput for many applications.
Nanotronics developed an extensive toolbox for computer vision and deep learning approaches to calculate spatial dimensions of imaged devices inline. We have expanded this technique for real-time measurement during video-faced focal search for critical depth measurement.
We use active template matching to compare and track the object through the stack to the known surface profile of the object. This allows for compensation of environmental noise, multi-object tracking within a single XY field-of-view and defect detection to be performed on the same images. Combined, this toolbox offers higher accuracy and faster processing than conventional approaches.
The Use of Deep Learning for Photometric Stereo of Noise Samples
Time: 4:20 PM - 4:45 PM
Authors: Denis Sharoukhov, Tina Narong, Vadim Pinskiy, Matthew Putman, Nanotronics (United States)
Photometric stereo is a technique for 3D reconstruction by calculating the surface normals of an object from different illumination angles. The technique is effective to estimate height profile of static objects with large features, but often fails for objects with smaller features or in flat environments with small depressions, like that of semiconductor wafers.
We explore compound methods of using neural networks, along with advancements in surface normal calculation, to compensate for expected noise. Nanotronics developed several methods of incorporating sample noise and uneven sample illumination in the photometric stereo model. They will demonstrate decreased noise susceptibility on synthetic and experimental datasets.
This approach enables the rapid inspection and reconstruction of complex surfaces without necessitating destructive or expensive analysis methods.
Nanotronics is a science technology company that has redefined Factory control through the invention of a platform that combines AI, automation and sophisticated imagining to assist human ingenuity in detecting flaws and anomalies in manufacturing, an industry that has been stagnant since the 1950’s.
Deployed across eight countries and industry agnostic, we work with leading-edge companies, from aerospace, to electronics, to healthcare, to drive up yield, reduce footprint and waste, lower costs, and speed up design iteration.
Nanotronics is a key player in helping to solidify New York’s role as a global center of the innovation economy. We are currently building New York’s first high tech manufacturing hub in the Brooklyn Navy Yard, with expected completion in Q2, 2020.To learn more visit https://nanotronics.co.
About nSpec® 3D
nSpec®3D combines advanced image-capture techniques with computational algorithms to provide surface topography, material roughness, and dispersion analysis. Tools can search, mark, and slice an image to identify features of interest.
About SPIE Photonics West
4-6 February, Moscone Center, San Francisco, California
The leading event for the photonics and laser communities with the latest components, devices, and systems for research or business needs.