This paper presents DBS-ElecNet, a deep learning framework for automated segmentation of electrodes and artifact regions in post-operative DBS MRI. To overcome reliance on manual annotations, we introduce a hybrid approach where a traditional image processing pipeline generates initial segmentations for the 3D U-Net model, which uses these as ground truth, and achieves robust segmentation performance. DBS-ElecNet performs inference in ~3 seconds, a 60-100x speedup over manual segmentations. This efficient and accurate approach enables scalable analysis for surgical verification and paves the way for advanced clinical applications like artifact inpainting.