Train a model to classify Day / Night images using keras.
Read full article here https://medium.com/@mneonizer/day-night-classification-a01a7d9af695
Requirements
pip install -r Requirements.txt
Training
python train_model.py --dataset images --model model/day_night.hd5
Images directory contains 2 sub-directories
├── images
│ ├── day
│ │ ├── 00001.jpg
│ │ ├── 00002.jpg
│ │ ├── 00003.jpg
│ │ ├── xxxxx.jpg
│ ├── night
│ │ ├── 00001.jpg
│ │ ├── 00002.jpg
│ │ ├── 00003.jpg
│ │ ├── xxxxx.jpg
And the model is trained to classify these two labels: day / night.
Testing
python test_model.py --model model/day_night.hd5 --image test/n2.jpg
python test_model.py --model model/day_night.hd5 --image test/d1.jpg
Reference: https://www.pyimagesearch.com/2017/12/11/image-classification-with-keras-and-deep-learning/