Enabling Efficient Deep Convolutional NeuralNetwork-based Sensor Fusion for AutonomousDriving
TimeTuesday, July 12th3:30pm - 3:54pm PDT
Location3000, Level 3
Event Type
Research Manuscript
Autonomous Systems (Automotive, Robotics, Drones)
Autonomous Systems
DescriptionMost automated vehicles are now equipped with multiple sensors enabling them to exploit complementary environmental context by fusing data from different sensing modalities. The fusion between DCNNs has been proved as a promising strategy to achieve satisfactory perception accuracy. However, existing DCNN fusion schemes conduct fusion by directly element-wisely adding feature maps extracted from different modalities together at various stages, failing to consider whether the features being fused are matched or not. We explore two feature-matching techniques. They enable DCNNto to learn corresponding feature maps with similar characteristics and complementary visual context from different modalities to achieve better accuracy.