- Kewen Zhang (kz2246@columbia.edu)
- Pengfei Wang (pw2406@columbia.edu)
- Yueying Teng (yt2495@columbia.edu)
This project is inspired by Kaggle.com "the-nature-conservancy-fisheries-monitoring" Link.
- Operating System: macOS Sierra
- Language: Python2.7
- Software: Pyspark, Opencv3
- detection_crop_train_fish.py: Crop fish in the training set
- detection_train_haar_classifier.py: Train Cascade classifier
- detection_generate_negative.py: Generate negative instances
- detection_clustering_ship.py: Cluster the fishery ships to denoise the raw images
- detection_crop_window.py: Get window of each fishery ships
- classification_get_sift_bow.py: Extract SIFT features and map into bag-of-words model
- classification_mllib_pyspark.py: Machine Learning with Pyspark
- classification_cnn.py: Construct Convolutional Neural Network model
- classification_cnn_LAG.py: Improve result by re-weighting