Computer Vision Approaches to Waste Containers Detection

@article{Valente2019ComputerVA,
  title={Computer Vision Approaches to Waste Containers Detection},
  author={M A Valente and H{\'e}lio Ricardo da Silva and Jo{\~a}o M. L. P. Caldeira and Vasco N. G. J. Soares and Pedro D. Gaspar},
  journal={2019 14th Iberian Conference on Information Systems and Technologies (CISTI)},
  year={2019},
  pages={1-4}
}
The work presented in this article is the result of a prelaminar investigation that aims at using computer vision techniques to replace the current method of performing detection of waste contains via radio-frequency identification. Comparatively to the current method, this approach is more agile and diminishes the resources needed for an implementation. The approach discussed is focused on the use of convolutional neural networks, specifically the network YOLO. Using this method of… CONTINUE READING

Results and Topics from this paper.

Key Quantitative Results

  • à iteração número 670 com uma perda média de 1.6 que não se encontrava perto de uma fase de estabilização.
  • Ainda assim, foi atingida uma precisão de deteção e classificação de 40% utilizando a última iteração da rede.
  • Verificou-se uma precisão máxima de identificação e classificação de 92% na segunda implementação da rede YOLOv2.

References

Publications referenced by this paper.
SHOWING 1-10 OF 12 REFERENCES

YOLO9000: Better, Faster, Stronger

  • 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2016
VIEW 11 EXCERPTS
HIGHLY INFLUENTIAL

YOLOv3: An Incremental Improvement

VIEW 7 EXCERPTS
HIGHLY INFLUENTIAL

darknet

A. Bezlepkin
  • GitHub repository. GitHub, 2019.
  • 2019
VIEW 1 EXCERPT

labelImg

T. Lin
  • GitHub repository. GitHub, 2019.
  • 2019
VIEW 1 EXCERPT

Object Detection with Deep Learning: A Review

  • IEEE transactions on neural networks and learning systems
  • 2018
VIEW 1 EXCERPT

You Only Look Once: Unified, Real-Time Object Detection

  • 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2015
VIEW 1 EXCERPT