A machine learning classification pipeline for detecting Lyman-alpha blobs in wide-field surveys using multi-band broadband data.
The pyBIA framework consists of four main modules
- catalog : Used to generate a catalog of morphological and intensity-based characteristics using image segmentation.
- ensemble_model : To train and optimize (including feature selection and hyperparameter tuning) a supervised learning classifier.
- outlier_detection : Used to extract features for image-based anomaly detection, and training an unsupervised anomaly detection algorithm.
- cnn_model : For processing multi-band imaging data (including pre-processing and data augmentation) and training a deep learning image classifier.
For more information including examples of how to use the code, please see the documentation page.
The latest stable version can be installed via pip.
$ pip install pyBIA
For technical details and an example of how to implement pyBIA, including how it was used in Godines & Prescott (2025), check out our Documentation.
Want to contribute? Bug detections? Comments? Suggestions? Please email us : [email protected], [email protected]