The submission of this project completed my Udacity Nanodegree: AWS Machine Learning Engineer. The assignment motivated me to make the most of my education. I combined my biochemistry and premed background into a coding project, reaching out to a professor in the biology department at Dartmouth to learn how I could apply my new skills to benefit his research.
This project is inspired by an issue that arises in cell biology, specifically, research on Drosophila embryos. The research project that this model aims to assist studies how genes work together to regulate cell behavior, thereby generating different tissue forms. Drosophila embryos are a great model for this research because of their genetic components, live-imaging accessibility and biophysical analysis. The relative ease of cell imaging allows researchers to photograph and classify a cell’s stage at different intervals.
In this research project, many images of individual cells were captured, manually classified, and further analyzed as per the research focus explained in the previous paragraph. However, classifying the embryonic stage for Drosophila cells is difficult, tedious and time consuming.
To expedite the research process, I have built an image classification model to classify the stage of embryonic cell development for Drosophila cell images. The project will be created entirely in AWS Sagemaker, using a Jupyter notebook and python training scripts. After manually uploading data to Sagemaker, the data will be augmented and uploaded to S3 via the Jupyter notebook.
Please see my Capstone Proposal and Project Report for more information on this project.