Abstract
In the present study we show a neural network based deductive experimental methodology for object recognition using CNN. Our goal is to generate a prototype which is based on feature mapping in combination with RPN and proposed log clipping using TNET for 3D detection, given by object recognition models of the KITTI platform ,focusing especially on AVOD and FPOINTNET, obtaining higher accuracy on smaller objects which are easily discarded by the point cloud provided by the 3d LIDAR HDL-64 laser sensor but not by feature mapping.
Translated title of the contribution | Analysis of Object Detection Algorithms for the Creation of a Prototype Based on the Fusion of Two Recognition Models |
---|---|
Original language | Spanish (Ecuador) |
Pages (from-to) | 5-10 |
Number of pages | 6 |
Journal | PRO Sciences |
Volume | 3 |
Issue number | 3 |
DOIs | |
State | Published - 29 Mar 2019 |
Keywords
- Cnn
- Frustun
- Kitti
- Lidar
CACES Knowledge Areas
- 116A Computer Science