Autonomous Object Detection
- 0 Collaborators
Object detection for autonomous navigation systems. This project provides core support for performing object detection on navigation datasets. Support for 3D object detection and domain adaptation are in experimental phase and will be added later. ...learn more
Project status: Published/In Market
Groups
Student Developers for AI
Intel Technologies
DevCloud,
Intel Python
Overview / Usage
This project is aimed at providing core support for performing object detection for autonomous navigation (in 2D/3D). It provides support for data distributions collected from structured/unstructured environments with incremental learning support (published in ICCVW 2019). It also provides support for augmentation, creating custom models. This makes it easier for researchers to perform rapid experimentation and experiment with different algorithms.
Methodology / Approach
To get started, there are 4 essential steps:
- Download the required dataset
- Setup dataset paths in
cfg.py - Create datalists
- Start training and evaluating/inference
It's that simple. If users are required to create their own models, they can easily do so in 4th step. We've provided an extensive documentation for how they can do it.
This project supports Pytorch>1.0, torchvision>=0.3
Technologies Used
-
Pytorch >= 1.0
-
Torchvision >= 0.3
-
Intel DevCloud