Week 6: Strategizing!

 Hello everyone,

   This week, I was conducting research into how to specifically go about creating data from the images. My first step will be to create code that automates creating a heat map of an image by combining morphological image processing and RGB coordinate data. The second step is to input the data into a generative adversarial neural network to determine whether the photo matches the training image sets provided to the algorithm. Automating the first step will be tremendously helpful as I have almost 20 different colonies for each bacteria type, and it would take forever to encode the photos myself. I hope to implement steps one and two together to make the algorithm as user-friendly as Google's teachable machine, yet more accurate as it's tuned to the specific purpose of identifying bacteria. But I'll do one step at a time first! This semester is flying by, so now that I've gotten a solid strategy, I will get started this weekend on coding. I hope everyone is having a good semester so far! 

My strategy for image processing:

Source: https://neptune.ai/blog/image-processing-python 

What my encoded images will look like to the algorithm:

Source: https://medium.com/analytics-vidhya/how-relevant-is-heatmap-in-your-machine-learning-model-6eb79a820f18

What the algorithm structure will look like: 

Source: https://neptune.ai/blog/image-processing-python