We need your help.

We have built a Convolutional Neural Network and need you to help train it. The goal for this phase of the project is to collect data from images labeled by geologists. Sedimentary structures can be subjective and we want a variety of opinions on each image to ensure we are getting the most accurate data for our model. These labeled images will then be ingested in our model to accurately classify different cross stratification in your data.

Label

Good models need good data. This is the grunt work of the operation. Hand labeling cross-stratification is the only way to ensure quality data in the long run. Our goal is to make this step as painless as possible.

Train

Once our test batch of images have been labeled, the data is used to train the model. We have developed a convolutional neural network model that will learn from the test images to classify new, unlabeled data.

Classify

Once the model is complete, we can then begin to apply it. Our future applications include processing drone footage, batch processing large databases, and expansion into planetary imagery.


Our team

Thumbnail Image

Keaton Cheffer
Geophysics Undergraduate

Keaton is a senior at Texas A&M University studying Geophysics and has a minor in Computer Science. His field of interest includes Deep Learning applications in Geoscience.

Thumbnail Image

Dr. Ryan Ewing
Associate Professor

Ryan’s research aims to understand the evolution of landscapes and the sedimentary record through physical processes operating at the surface-atmosphere interface of Earth, Mars and Titan.



Work with us