Update README.md
Browse files
README.md
CHANGED
@@ -12,7 +12,8 @@ license: mit
|
|
12 |
|
13 |
# Enhancing Visual Place Recognition via Fast and Slow Adaptive Biasing in Event Cameras
|
14 |
|
15 |
-
[](https://www.youtube.com/watch?v=8D9gtHqteEQ)[](https://gokulbnr.github.io/publication/dvs-biasing-vpr)[](https://github.com/gokulbnr/fast-slow-biased-event-vpr)[](https://arxiv.org/abs/2403.16425)[, to be presented at the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024). This work introduces feedback control algorithms that dynamically change bias parameters for event-cameras to stabilize event-rate in an online manner. The work reports improvements in visual place recognition performances across variations in environment brightness conditions, validated through comprehensive real-time evaluations using a new [QCR-Fast-and-Slow-Event-Dataset](https://huggingface.co/datasets/gokulbnr/QCR-Fast-Slow-Event-Dataset).
|
18 |
|
|
|
12 |
|
13 |
# Enhancing Visual Place Recognition via Fast and Slow Adaptive Biasing in Event Cameras
|
14 |
|
15 |
+
[](https://www.youtube.com/watch?v=8D9gtHqteEQ)[](https://gokulbnr.github.io/publication/dvs-biasing-vpr)[](https://github.com/gokulbnr/fast-slow-biased-event-vpr)[](https://arxiv.org/abs/2403.16425)[](https://ieeexplore.ieee.org/document/10802384)
|
16 |
+
|
17 |
|
18 |
Welcome to the official **QCR-Fast-Slow-Event-Dataset** dataset repository attached to the paper [**Enhancing Visual Place Recognition via Fast and Slow Adaptive Biasing in Event Cameras**](https://arxiv.org/abs/2403.16425), to be presented at the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024). This work introduces feedback control algorithms that dynamically change bias parameters for event-cameras to stabilize event-rate in an online manner. The work reports improvements in visual place recognition performances across variations in environment brightness conditions, validated through comprehensive real-time evaluations using a new [QCR-Fast-and-Slow-Event-Dataset](https://huggingface.co/datasets/gokulbnr/QCR-Fast-Slow-Event-Dataset).
|
19 |
|