Data processing stuff for a bunch of fire videos we got

HOW TO USE

We have two interactive options:

Simple neural network: uses k-means clustering to give the time in a video a block is expected to catch fire, or if it will catch fire at all.

Time prediction neural network: predicts how long it will be until a block catches fire.

Simple neural network

1. Put an image in the image_simple directory.
2. Run test_nn.py (also runs image_extractor.py).

Time prediction neural network

1. Put an image in the predict_time_josh/image directory.
2. Run Predictor.py.

FILE EXPLANATION

Not everything included in tree, some unimportant files are not included

CS_429_Final_Project
├── Potential_Deliverables/ (DIRECTORY: Charts from data analysis varying wind speed and block size weight)
│
├── predict_time_josh/  (DIRECTORY: used for prediction model)
│   │
│   ├── FIND_BEST_MODEL/  (DIRECTORY: used to find best model)
│   │  │
│   │  ├── ####_iteration/  (DIRECTORIES: Scripts, models, and text outputs from 7 iterations)
│   │  │
│   │  ├── further_training/  (DIRECTORY: Scripts and models for fine-tuning)
│   │  │
│   │  ├── all_archs.txt  (All architectures tested)
│   │  │
│   │  └── PCA_20_flame_frame_data.csv (Training data)
│   │
│   ├── image/  (DIRECTORY: Image to be predicted)
│   │
│   ├── old_data/  (DIRECTORY: Old data not used but kept)
│   │
│   │
│   ├── pca_model.pkl (PCA model for predictions)
│   │
│   ├── PREDICTOR.py (Final model prediction script)
│   │
│   └── separate_features.py (Separates video data into features/labels)
│
│
├── image_simple/  (DIRECTORY: for images for simple neural network)
│
│
├── image_extractor.py  (Converts image to feature vector CSV)
│
├── image_output.csv  (Output of image_extractor.py)
│
├── data_preprocess.py  (Main preprocessing script)
│
├── change_class_info.py  (Script to alter wind speed and block size weight - unused)
│
├── data_pca.py  (Deprecated PCA script)
│
├── k_means.py  (Runs k-means clustering)
│
├── convenience.py  (Convenience script for data collection)
│
├── simple_nn.py  (Simple neural network for classification)
│
└── test_nn.py  (Predicts fire timing and status)