2020 BRN Discussion, page-340

I don't know if you PPL have seem the lastest by Shibo Zuo aka Bob from BRN

https://www.linkedin.com/in/bob-zhou/

• A Spike Learning System for Event-driven Object Recognition
  
  publication date Dec 31, 2019   publication description IJCV
• publication description Computer vision tasks like object recognition and detection have achieved great success with the booming of deep neural networks (DNNs). However, the fixed frame
  rate of traditional cameras has set tough challenges on the road ahead, causing
  severe problems such as motion blur and high computational overhead in applications
  with high speed and dynamics.
• Event-driven sensors such as LiDAR and dynamic
  vision sensor (DVS) have been developed to acquire better sensing qualities in these
  scenarios, whereas most existing methods heavily rely on the pre-processing of the
  pulse signals derived from such sensors, and therefore result in considerable
  computational and time cost.
• In this paper, we investigated the spiking neural network
  (SNN) an approach utilizing temporal coding to address the object recognition problem
  directly with raw temporal pulses.
• To help with the evaluation and benchmarking,
  experiments were performed on both LiDAR and DVS datasets, including KITTI object
  detection dataset, a self-created comprehensive temporal pulses dataset simulating
  LiDAR reflection in different road scenarios, DVS-barrel dataset, and DVS-CIFAR10
  dataset.
• Being evaluated on different datasets, our proposed method has shown state-of-the-art performance while achieving remarkable time efficiency. It highlights the
  potential of SNN in challenging applications with event-driven sensors.

Deep SCNN-based Real-time Object Detection for Self-driving Vehicles Using LiDAR Temporal Data

publication date Dec 30, 2019   publication description IEEE

publication description Real-time accurate detection of three-dimensional (3D) objects is a fundamental necessity for self-driving vehicles.
Traditional computer-vision approaches are based on convolutional neural networks (CNN). Although the accuracy of using CNN on the KITTI vision benchmark dataset has resulted in great success, few related studies have examined its energy consumption requirements.
Spiking neural networks (SNN) and spiking-CNNs (SCNN) have exhibited lower energy consumption rates than CNN. However, few studies have used SNNs or SCNNs to detect objects.
Therefore, we developed a novel data preprocessing layer that translates 3D point-cloud spike times into input and employs SCNN on a YOLOv2 architecture to detect objects via spiking signals. Moreover, we present an estimation method for energy consumption and network sparsity.
The results demonstrate that the proposed networks ran with a much higher frame rate of 35.7 fps on an NVIDIA GTX 1080i graphical processing unit.
Additionally, the proposed networks with skip connections showed better performance than those without skip connections. Both reached state-of-the-art detection accuracy on the KITTI dataset, and our networks consumed an average (low) energy of 0.585 mJ with a mean sparsity of 56.24%.

And also

https://deepai.org/publication/deep...lf-driving-vehicles-using-lidar-temporal-data