1. Home

MACHINE LEARNING FOR HUMAN RECOGNITION

Deņiss Surikovs, Andrejs Mežārs, Ritvars Bleive, Sergejs Kodors

Abstract


The aim of this work is to develop a neural network that will be able to recognize human presence. To achieve this goal, authors applied neural network architecture YOLOv5 and the open dataset COCO. The experiment was repeated three times. The study yielded a good result - the neural network was able to detect humans in images with a good precision.

Keywords


artificial neural network; COCO dataset; human recognition; YOLOv5

Full Text:

PDF

References


Human detection and tracking with deep convolutional neural networks [tiešsaiste], [atsauce uz 01.05.2022]. Pieejams: https://link.springer.com/article/10.1007/s11042-020-09579-x

Konvolūcijas neironu tīkls [tiešsaiste], [atsauce uz 01.05.2022]. Pieejams: https://www.techtarget.com/searchenterpriseai/definition/convolutional-neural-network

YOLOv5: The Latest Model for Object Detection [tiešsaiste], [atsauce uz 01.05.2022]. Pieejams: https://medium.com/axinc-ai/yolov5-the-latest-model-for-object-detection-b13320ec516b

Comon Objects In Context [tiešsaiste], [atsauce uz 09.05.2022]. Pieejams: https://cocodataset.org/#download

FiftyOne [tiešsaiste], [atsauce uz 10.05.2022]. Pieejams: https://voxel51.com/fiftyone/index.html

Computer Vision Datasets [tiešsaiste], [atsauce uz 09.05.2022]. Pieejams: https://public.roboflow.com/




DOI: https://doi.org/10.17770/het2022.26.6956

Refbacks

  • There are currently no refbacks.
Rēzeknes Tehnoloģiju akadēmija, 2016 © Atbrīvošanas aleja 115, Rēzekne, Latvija. Tālrunis: 28325368;
Jautājumus un ierosinājumus sūtiet uz helpdesk@rta.lv