A deep neural network approach to compact source removal

Context. Analyzing extended emission in photometric observations of star-forming regions requires maps free from compact foreground, embedded, and background sources, which can interfere with various techniques used to characterize the interstellar medium. Within the framework of the NEMESIS project, we applied machine-learning techniques to improve our understanding of the star formation timescales, which involves the unbiased analysis of the extended emission in these regions. Aims. We present a deep learning-based method for separating the signals of compact sources and extended emission in photometric observations made by the Herschel Space Observatory, facilitating the analysis of extended emission and improving the photometry of compact sources. Methods. Central to our approach is a modified U-Net architecture with partial convolutional layers. This method enables effective source removal and background estimation across various flux densities, using a series of partial convolutional layers, batch normalization, and ReLU activation layers within blocks. Our training process utilized simulated sources injected into Herschel images, with controlled flux densities against known backgrounds. A dynamic, signal-to-noise ratio (S/N)-based adaptive masking system was implemented to assess how prominently a compact source stands out from the surrounding background. Results. The results demonstrate that our method can significantly improve the photometric accuracy in the presence of highly fluctuating backgrounds. Moreover, the approach can preserve all characteristics of the images, including the noise properties. Conclusions. The presented approach allows users to analyze extended emission without the interference of disturbing point sources or perform more precise photometry of sources located in complex environments. We also provide a Python tool with tutorials and examples to help the community effectively utilize this method.