LithographyQuality

Analyzing the quality of a nanofabricated lithography pattern.

To compute the quality of the lithography pattern, you can use a combination of techniques like image registration and structural similarity index measure (SSIM). Image registration will help you align the generated pattern with the original pattern, and SSIM will provide you with a quality score.

--> https://en.wikipedia.org/wiki/Structural_similarity#:~:text=SSIM%20is%20used%20for%20measuring,distortion%2Dfree%20image%20as%20reference.

Goals:

Pre-Process : Trim Edges and Axes from Scan Image. Convert Lithography Scan to Grey-Scale to SSIM with original Pattern. SSIM = Quality (0-1 Value, 0 Being Worst, 1 Being Perfect)

1. Iterate overlays for multiple sub-sections of a larger lithography array.

2. Plot the quality score as a multivariable function of input lithography parameters

3. Find optimal score within a error domain of input parameters Incorporate Principal Component Analysis (PCA) as there may be too many variables to adequatly visualize... https://builtin.com/data-science/step-step-explanation-principal-component-analysis --> Embedding : https://www.toptal.com/machine-learning/embeddings-in-machine-learning#:~:text=An%20embedding%20is%20a%20low,for%20a%20particular%20data%20structure.

4. Do the same with final post-LiftOff patterns and compare quality scores.