Deep Transfer Learning for Blended Source Identification in Galaxy Survey Data

We present BlendHunter, a proof-of-concept for a deep transfer learning based approach for the automated and robust identification of blended sources in galaxy survey data. We take the VGG-16 network with pre-trained convolutional layers and train the fully connected layers on parametric models of COSMOS images. We test the efficacy of the transfer learning by taking the weights learned on the parametric models and using them to identify blends in more realistic CFIS-like images. We compare the performance of this method to SEP (a Python implementation of SExtractor) as function of noise level and the separation between sources. We find that BlendHunter outperforms SEP by $\sim 15\%$ in terms of classification accuracy for close blends ($<10$ pixel separation between sources) regardless of the noise level used for training. Additionally, the method provides consistent results to SEP for distant blends ($\geq10$ pixel separation between sources) provided the network is trained on data with a relatively close noise standard deviation to the target images. The code and data have been made publicly available to ensure the reproducibility of the results.