ASML
Graduation assignment: Investigate image feature extraction and classification via deep learning
Introduction
This is a graduation assignment for a master student with a background in Physics, Mathematics, Electrical Engineering or related with a focus on Image Processing, (Mathematical) Optimization and Machine Learning.
Job Mission
Within the Development & Engineering department at ASML, the group YieldStar Algorithms and Image Processing covers the development of physical, optical and mathematical models and methods. These are required to infer Overlay and Focus from scatterometry data. Relevant new metrics and algorithms, as well as new measurement functions, with optimum performance characteristics are identified, designed and implemented. The group is also responsible for contributing to the invention of new ASML products providing state-of-the-art expertise in Scatterometry, Machine Learning, Mathematical Optimization, Image and Signal Processing.
Chips continue to get faster and smaller -- and more difficult to be manufactured. ASML's Holistic Lithography products help chipmakers to squeeze every bit of performance out of their lithography equipment. An important part of the Holistic Lithography package is YieldStar, a metrology tool that gathers data to maximize the performance of a lithography system -- which in turn allows manufacturers to produce more advanced chips. YieldStar works fast and with high accuracy because it uses an advanced technology called scatterometry..
The goal of this study is to build up a neural network to extract discriminative features from the acquired data. Instead of manually extracting data features, input data will be directly fed into a deep neural network that learns the features automatically. As part of the work, also the design of an image classifier that uses the extracted features and a training dataset of images will be carried out.