| Title |
Efficiency of YOLOv5 models in the detection of construction details / |
| Authors |
Kvietkauskas, Tautvydas ; Pavlov, Ernest ; Stefanovič, Pavel |
| DOI |
10.15388/DAMSS.14.2023 |
| ISBN |
9786090709856 |
| Full Text |
|
| Is Part of |
DAMSS 2023: 14th conference on data analysis methods for software systems, Druskininkai, Lithuania, November 30 - December 2, 2023.. Vilnius : Vilniaus universiteto leidykla, 2023. p. 46.. ISBN 9786090709856 |
| Keywords [eng] |
object detectiom ; contruction details ; YOLOv5 |
| Abstract [eng] |
Object detection today is widely used in different areas, for example, medicine, industry, business, and even everyday solutions. New models of object detection are constantly developed and the old models are improved by adding some new features or changing the architecture of models. One of the most used object detection models in scientific research is YOLO. In recent years, some new versions of YOLO have been proposed, but they are not fully investigated and lack scientific research. The most stable version of the YOLO group algorithm is the YOLOv5. In this research, the newly collected construction detail dataset has been prepared. Images from different angles of construction details have been taken, and the dataset has been labelled. The dataset consists of 22 construction details. The experimental investigation has been performed using five different models of YOLOv5: n, s, m, l, x. During the experimental investigation, various parameters have been used to find out the influence of the parameters on the final detection results. The models have been tested on three different backgrounds: white, neutral, and mixed. The results of the experimental investigation are promising, and in the future, the models can be used in construction recommendation models. |
| Published |
Vilnius : Vilniaus universiteto leidykla, 2023 |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2023 |
| CC license |
|
