 The paper proposes a new approach for detecting overshooting tops using deep learning, specifically a convolutional neural network, CNN, which can accurately identify OTs through visual inspection of contextual information in Himawari 8 satellite images. The validation results show that the CNN can successfully detect OTs over tropical regions with a mean probability of detection, POD, of 79.68% and a mean false alarm ratio, FAH, of 9.78%. This article was authored by Maikim, Jung-Hae Lee, and Jung-Ho Im.