https://github.com/lucaangioloni/objecttrajectories

Estimate object trajectories in an automotive context, using KITTI dataset, with CNNs.
https://github.com/lucaangioloni/objecttrajectories

automotive cnn kitti-dataset trajectories vision

Last synced: 2 months ago
JSON representation

Estimate object trajectories in an automotive context, using KITTI dataset, with CNNs.

• Host: GitHub
• URL: https://github.com/lucaangioloni/objecttrajectories
• Owner: LucaAngioloni
• License: mit
• Created: 2017-06-30T16:09:23.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2018-01-25T16:44:26.000Z (over 7 years ago)
• Last Synced: 2025-04-23T18:47:36.869Z (6 months ago)
• Topics: automotive, cnn, kitti-dataset, trajectories, vision
• Language: Jupyter Notebook
• Size: 7.93 MB
• Stars: 8
• Watchers: 2
• Forks: 4
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# ObjectTrajectories

Estimate object trajectories in an automotive context, using [KITTI dataset](http://www.cvlibs.net/datasets/kitti/index.php), with CNNs.

A very detailed description of the implamentation is available in the `relazione.pdf` file (but in italian).

![TestImage1](./images/consistenza.png)

![TestImage2](./images/occlusioni.png)

The project dependencies are the following Python packets (modules):

- Pillow

- Numpy

- GeoPy

- Keras

- TensorFlow

The CNNs code is based on the following GitHub repos:

- [YOLO: YAD2K - Yet Another Darknet 2 Keras](https://github.com/allanzelener/YAD2K)

- [SSD: Single Shot MultiBox Detector in Keras](https://github.com/oarriaga/single_shot_multibox_detector)