The rapid development of embedded hardware in autonomous vehicles broadens\ntheir computational capabilities, thus bringing the possibility to mount more\ncomplete sensor setups able to handle driving scenarios of higher complexity.\nAs a result, new challenges such as multiple detections of the same object have\nto be addressed. In this work, a siamese network is integrated into the\npipeline of a well-known 3D object detector approach to suppress duplicate\nproposals coming from different cameras via re-identification. Additionally,\nassociations are exploited to enhance the 3D box regression of the object by\naggregating their corresponding LiDAR frustums. The experimental evaluation on\nthe nuScenes dataset shows that the proposed method outperforms traditional NMS\napproaches.\n