Browsing by Subject "Learning techniques"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Determination of tribological properties at CuSn10 alloy journal bearings by experimental and means of artificial neural networks method
    (2012) Ünlü B.S.; Durmuş H.; Meriç C.
    Purpose - It is important to know the friction coefficient and wear loss for determination of tribological conditions at journal bearings. Tribological events that influence wear and its variations affect experimental results. The purpose of this paper is to determine friction coefficient and wear loss at CuSn10 alloy radial bearings by a new approach. In experiments, effects of bearings have been examined at dry and lubricated conditions and at different loads and velocities. Design/methodology/approach - In this study, friction coefficient and wear losses of journal and bearing have been determined by a new approach with a radial journal bearing test rig and artificial neural networks (ANNs) method. The ANN typifies a learning technique that enables the hidden input-output relationship to be mapped accurately. Bronze-based materials have been used as bearing material. Effects of friction coefficient and wear losses have been examined at same load and velocity and at dry and lubricated conditions. Findings - The results obtained in ANN application are close to friction test results for dry and lubricated conditions. Therefore, by using trained ANN values, the intermediate results that were not obtained in the tests can be calculated. Experimental studies will be increased and research with ANN will be continued. Originality/value - By using trained ANN values, the intermediate results that were not obtained in the tests can be calculated. The training finished on 30 min whereas experimental study had continued day after day. Copyright © 2012 Emerald Group Publishing Limited. All rights reserved.
  • No Thumbnail Available
    Item
    A machine learning framework for full-reference 3D shape quality assessment
    (Springer, 2020) Yildiz Z.C.; Oztireli A.C.; Capin T.
    To decide whether the perceived quality of a mesh is influenced by a certain modification such as compression or simplification, a metric for estimating the visual quality of 3D meshes is required. Today, machine learning and deep learning techniques are getting increasingly popular since they present efficient solutions to many complex problems. However, these techniques are not much utilized in the field of 3D shape perception. We propose a novel machine learning-based approach for evaluating the visual quality of 3D static meshes. The novelty of our study lies in incorporating crowdsourcing in a machine learning framework for visual quality evaluation. We deliberate that this is an elegant way since modeling human visual system processes is a tedious task and requires tuning many parameters. We employ crowdsourcing methodology for collecting data of quality evaluations and metric learning for drawing the best parameters that well correlate with the human perception. Experimental validation of the proposed metric reveals a promising correlation between the metric output and human perception. Results of our crowdsourcing experiments are publicly available for the community. © 2018, Springer-Verlag GmbH Germany, part of Springer Nature.
  • No Thumbnail Available
    Item
    HeatWatch: Identification of Parameters Influencing the Urban Heat Island Effect through Deep Learning Techniques
    (International Society for Photogrammetry and Remote Sensing, 2024) Kılınç M.; Aydın C.; Aydın G.E.; Balcı D.
    Global climate change (GCC) is accelerated by factors such as greenhouse gas emissions from human activities and the urban heat island (UHI) effect, particularly in densely urbanized areas. According to the WMO's 2023 data, GCC warming effects increased by 49% from 1990s to 2021, 80% of this increase was due to CO2. Furthermore, average global temperatures have increased by 1.1°C since the early 1900s. In this respect, the urban heat island (UHI) effect has gained importance with global temperatures. Environmental problems that cause cities to be warmer than rural areas, especially due to hard ground surfaces, building density and decreasing green areas, have the potential to create negative impacts on human health. Therefore, it is important to identify and manage the impact of UHI. This is because traditional methods are limited to fixed station data, but technologies such as remote sensing and geographic information systems (GIS) provide more comprehensive results for this management. In an innovative approach, deep learning and artificial intelligence techniques can provide more accurate analysis by processing large datasets. In this context, this research proposes a conceptual framework for using deep learning techniques to detect the UHI effect with data obtained from street images and unmanned aerial vehicles. With the proposal, the UHI value will be calculated by detecting objects such as trees, air conditioners, vehicles and building cladding with the YOLO-Real-Time Object Detection algorithm. With this approach, it is aimed to obtaining more precise and accurate results in determining the UHI effect. In addition, a web-based management panel will be designed for managers to review the results and use them in decision-making mechanisms. It is aimed at disseminating this model and making it an important tool in the planning of urban areas. © Author(s) 2024. CC BY 4.0 License.