BEV-MODNet: Monocular Camera based Bird's Eye View Moving Object Detection for Autonomous Driving

Exploiting Multi-Modal Fusion for Urban Autonomous Driving Using Latent Deep Reinforcement Learning

“…Such architectures face difficulty adapting to challenging situations as their designs are complex enough and are created exclusively for specific challenging scenarios. Recent advancements in learning-based approaches lead to massive progress in motion detection [161,162,163,93,157,42,164,165]. SMSnet [161] is a method that leverages a convolutional neural network (CNN) and depends on two sequential camera images to perform pixel-wise category labeling and motion detection.…”

“…Unlike the earlier versions, [164] improves the diversity of moving objects by adding four additional classes instead of just vehicles. Lastly, BEV-MODNet [165] is another enhancement that investigated the idea of learning motion detections directly on the BEV space. In [165], a deep network is designed with a two-stream RGB and an optical flow fusion architecture.…”

LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction

“…Such architectures face difficulty adapting to challenging situations as their designs are complex enough and are created exclusively for specific challenging scenarios. Recent advancements in learning-based approaches lead to massive progress in motion detection [8], [12], [13], [18]- [22]. SMSnet [18] is a method that leverages a convolutional neural network (CNN) and depends on two sequential camera images to perform pixel-wise category labeling and motion detection.…”

“…Unlike the earlier versions, [21] improves the diversity of moving objects by adding four additional classes instead of just vehicles. Lastly, BEV-MODNet [22] is another enhancement that investigated the idea of learning motion detections directly on the BEV space. In [22], a deep network is designed with a two-stream RGB and an optical flow fusion architecture.…”

“…Lastly, BEV-MODNet [22] is another enhancement that investigated the idea of learning motion detections directly on the BEV space. In [22], a deep network is designed with a two-stream RGB and an optical flow fusion architecture.…”