At this point in the semester, we need to start obtaining some results, even if we know that they are not near the best ones we can obtain. We have 2 segmentation methods based on thresholding and edge detection that show decent accuracy in identifying the boundaries. The following schedule will be adopted for the rest of the term:

While exploring ways to improve segmentation in the background:

1) Obtain optical images of the chips resulting from the use of the following tool parameters.

2) For each case, obtain images at 3 approximate areas from 10 chips. Namely (as shown in the previous post):

• Left Shear Zone
• Right Shear Zone
• Middle Region

3) Create 2 separate data sets using each segmentation method.

4) Divide shear zones into subsections based on their distance from the edges as necessary. This will be resorted to as often as the analysis at later stages warrants it.

5) Calculate 2 Point Statistics on each of the data sets.

6) Cast the data sets to the principal component space.

7) Try to identify patterns correlations in the PC space that correspond to process parameters and the distance from the edges of the individual chips.