Interactive Machine Learning Project to ID Stairs
Stan Rosenbaum - Fall 2017
Environment:
- Anaconda 5.0 / Python 3.6
- TensorFlow 1.1.0
- Keras 2.0.8
The Main Scripts are probably:
* CS-660-SP-Notebook-1-V2.ipynb - More of a scratch pad.
* HumanOracle.py - The Human Oracle Script
* NNExploration.py - For Building and Manipulation of the Keras CNN
* CS660DataManagement.py - For Manipulating and Handling the Data
Folder Directory Structure:
\Anaconda - Jupyter Notebook and Python scripts
\Data - Collected Images and Point Cloud Data
\Data\Combined - All Collected data sorted into data flavors - The NORM and FLOP examples combined. '0' at end of name is NORM, '1' is FLOP.
\Data\Pickles - Python Pickles of the data it uses
\Data\Raw - The Raw Data
\Model_Structs - Saved Keras Models
\Notes - General Notes [like this]
\Results - Results from Training and Test Analysis, A Excel and CSV files mostly.
\Tensor_Log - For TensorBoard
\Tensor_Logs_Archive - Archive for above.