About me


Current Resume (pdf)

I create real-time computer vision based applications implemented in C++ and Python. My areas of work include Image/Video Analytics, Non-Rigid Medical Image Registration, Radar, and Lidar using Deep Learning and Machine Learning. Systems include GPU/Cuda, Grid Engine computing, Matlab, Python, AWS, and Azure.

I got my PhD from the ECE School at Georgia Tech in 2007.  There I worked for Professor Anthony Yezzi where I designed and implemented vision systems for segmentation, non-rigid registration, and tracking using Curve Evolution through PDE’s with C++ and Tcl/Tk.

My github is here

Some of my github repos:

  1. Deep Learning (repo)
    1. Getting Tensorflow & PyTorch to work on Win10 using Anaconda Python (folder)
    2. Onnx (folder)
    3. Tensorflow Examples (folder)
    4. PyTorch Tutorials (folder)
  2. OpenCV Ipython Notebooks (folder)
  3. Modded IPython Notebooks (folder)
  4. Games AI
    1. Desmume Emulator with Lua for Nintendo DS (repo)
  5. Geospatial Jupyter Notebooks (repo)
  6. ITorch Notebook
    1. Logistic Regression ipynb
  7. Support Files for CatholicVideoBible (repo)
    1. Instagram, YouTube, TikTok
    2. Produces CatholicVideoBibleData (repo)
  8. Support Files for DeepFakeCeleb (repo)
    1. YouTube, TikTok
  9. Game Translations (repo)
  10. Bash shell examples (repo)
  11. Quantitative Finance with quantecon (repo)
  12. Lots of Notes on many subjects (repo)
    1. Deep Learning (folder)
    2. Filtering (folder)
    3. Geometry (folder)
    4. Geospatial (folder)
    5. Interviews (folder)
    6. Machine Learning (folder)
      1. ICA (folder)
    7. Quantitative Finance (folder)
    8. Radar (folder)
  13. Caffe
    1. Train Cifar10 using CPP Caffe: ipynb
    2. Test Random Image from Google using Caffe Trained Cifar10 network: ipynb
  14. Docker (dockerfiles directory on github)
    1. Cuda: github, dockerhub
    2. Caffe+Cuda: github, dockerhub
    3. IpythonNotebook+Caffe+Cuda: github

Flipboard news is here

Leave a Reply

Your email address will not be published.