Researchers have developed a network that combines 3D LIDAR and 2D image data to enhance the detection of small objects in 3D space, a significant step forward for robots and self-driving cars.

However, the accurate detection of small objects remains a challenge.

The MPPF module, which is part of this research, is tasked with performing explicit interactions between intra-modal and cross-modal features.