[CV 2018 / Paper Summary ]Complex-YOLO: An Euler-Region-Proposal for Real-time 3D Object Detection on Point Clouds

3D-Object Detection
  1. Processing the point cloud directly using multi layer perceptron
  2. Translating the point cloud data into voxel regions and applying CNN
  3. Combination/ Fusion of the Approach 1 & 2
ERPN values
Anchor boxes
3D bbox regressor
Loss Function
  1. The ERPN is a part of the yolo network
  2. One end to end training as its a single network
  3. Prediction of bounding boxes happens in one single inference pass
  4. Lower runtime in comparision wrt to other models having RPN
Design efficiency
  1. 2D object detection
  2. 3D object detection
  3. Birds eye view
  1. Optimizer: SGD
  2. Decay:0.0005
  3. Momentum : 0.9
  4. Training set:85% of data
  5. Validation set:15% of data
  6. Epochs:1000
  7. Learning rate: small at the beginning then gradually increased
  8. Activation fuction : Leaky Relu
  9. Regularization: Batch Norm
  1. IOU threshold :0.7 Car,0.5 Pedestrian and Cyclist
  2. Metric: Average Precission (AP)
  1. Birds Eye View : Complex yolo is able to achieve good results in comparission with other state of the art algorithm , the positive point is it is able to achieve high FPS on a Titan X gpu compared to other architecture
Performance Results of Birds eye view
Performance Comparision of 3D object detection
Peformance plot

Final Words ….


  1. Redmon, J.: Darknet: Open source neural networks in c. http://pjreddie.com/darknet/ (2013–2016)
  2. Chen, X., Kundu, K., Zhu, Y., Berneshawi, A., Ma, H., Fidler, S., Urtasun, R.: 3dobject proposals for accurate object class detection. In: NIPS. (2015)
  3. Redmon, J., Farhadi, A.:YOLO9000: better, faster, stronger.CoRR
    abs/1612.08242 (2016)
  4. Chen, X., Ma, H., Wan, J., Li, B., Xia, T.: Multi-view 3d object detection network for autonomous driving. CoRR abs/1611.07759 (2016)



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store