The goal of this proposal is to create the software Weave to automate the extraction of critical dimensions (CDs) from scanning electron microscope (SEM) images for the microelectronics industry. Current best practices for extraction of CDs are that personnel analyze the images one by one, which is tedious, prone to human bias, time-consuming and expensive. Successful implementation of Weave will save almost 10% of process engineers time, freeing them for more productive and creative work. Two methods will be developed, tested and compared for edge and material detection and automated extraction of CDs; level set (LS) contour or edge detection and machine learning (ML) edge detection. Both methods will be developed because both show promise for this application, and they need to be evaluated for speed and accuracy on a variety of typical device nanostructures. The software will allow the user to import a prototypical image and specify the measurements to be extracted. The user then submits all the images with similar features (e.g., trenches or finfets) to be analyzed, and the software will produce tables of CD measurements and annotated images. Weave will plug directly into our current SandBox Studio software tool providing semiconductor process engineers with an end-to-end recipe development solution.