OctreeNet_CRF details

About

Short Name OctreeNet_CRF
Long Name OctreeNet_CRF
Website
Description A novel sparse convolutional neural network based on the octree for real-time point cloud classification. The proposed framework only uses the coordinates of 3D points which are direct outputs of standard laser scanners and no color information is involved. The submitted runtime includes the io time. The time costs for io and data construction, OctreeNet inference and CRF inference are 172.57s, 5.93s and 6.34s respectively.
Reference F. Wang, Y. Zhuang, H. Gu, and H. Hu, OctreeNet:A Novel Sparse 3D Convolutional Neural Network for Real-time 3D Outdoor Scene Analysis, submitted to IEEE Transactions on Automation Science and Engineering.
Hardware Intel® core(TM) CPU i5-4590 @ 3.3 GHZ, 8 GB RAM
Used additional training data 0
Last submission 2017-12-13 07:26:36
Is opensource 0
Number of submissions 1

Result

A_IoU OA [s] IoU 1 IoU 2 IoU 3 IoU 4 IoU 5 IoU 6 IoU 7 IoU 8
Result0.5910.899184.840.9070.8200.8240.3930.9000.1090.3120.460
Rank32434334957

Full Confusion Matrix

L1' L2' L3' L4' L5' L6' L7' L8'
L1138745792694371449114031309391975862363708
L2592931984792214798217248106211104111966
L36123294597948118105525066505731250
L41064192425307798931431536151454226531746588873
L59803662397364664169975296487172131914247150179
L6857771352274137314739814887723155508639384055
L710332618146733535133619836146062835
L89031134293475620751580052340123250460132