Room: Old Main Academic Center 1030
Chair: Shiraz Mujahid
Sarah Stirrat, University of Strathclyde, Scotland
Time: 2:40 pm - 3:00 pm (CST)
Title: Mapping the parameter space of trailing-edge noise – aero acoustic modelling and optimization
Abstract:
Aircraft noise can be split into that associated with the engine or the airframe. Engine associated noise is generated from the internal moving surfaces within the engine components as well as from the exhaust gas emanating at the nozzle exit. The breakdown of the latter results in both jet noise and jet-surface interaction noise components when the turbulent air interacts with the airframe, wing edges and other external surfaces. The trailing-edge component is a particularly dangerous noise source owing to the large increase in low frequency sound when the observation point is above or below the plate surface and vertical location (h) of the trailing-edge is of the order of jet diameter.
In previous work we investigated the performance of various numerical algorithms to find the optimum mean flow and turbulence parameters for the acoustic spectrum of a trailing-edge noise model. This optimum was determined to minimize an appropriate acoustic norm, i.e., proportional to the difference between the acoustic data and that predicted by the aero-acoustic model for a particular jet/plate/acoustic observation point configuration.
In this presentation, we explain this approach and show its robustness in predicting trailing edge noise over a parameter range that includes jet acoustic Mach number, (streamwise, vertical) location of the trailing edge as well as the polar angle, , measured from the jet axis for a particular jet/plate configuration.
(Joint work with M. Z. A. Koshuriyan)
Ciprian Orhei, Politehnica University of Timisoara
Time: 3:00 pm - 3:20 pm (CST)
Title: Semantic aware urban landmark detection system
Abstract:
Landmarks in an urban area serve as “spatial magnet” in which cultural, civic, or economical activities take place. In this sense they have become an important aspect in multiple domains related to tourism and culture. The proposed algorithm, which is an extension of our previous work [Orhei et al., 2021], is capable of handling difficult situations or scenarios when multiple landmarks of interest are clustered into a small arial like city squares or tourist neighbourhoods. The offline path takes a dataset of images and generates vectors of features that are refined using a semantic segmentation-based region of interest. Afterwards, the features vectors are grouped together in a vector for each landmark. In the on-line phase we mirror the pre-processing part and using the classifier we find the closest similar landmark vector to the inquiry image. To enhance the detection, we use GPS tags to limit our search range within the clusters. By doing so we gain benefits in the direction of detection accuracy and run-time. The proposed system has obtained a 99.00% detection rate on the popular ZuBuD dataset and 92.05% on TMBuD dataset.
(Joint work with Radu Vasiu)
Madalina Sofia Pasca, Politehnica University of Timisoara and West University Timisoara
Time: 3:40 pm - 4:00 pm (CST)
Title: Analytical approximate solution for the first order nonlinear differential equations using Piecewise Polynomial Least Squares Method
Abstract:
In this paper I present a new method for determining approximate analytical solutions for di erential equations, the Piecewise Polynomial Least Squares Method (PWPLSM). In the attached numerical examples you can see the ease of application of the method as well as its accuracy. I also present comparisons with the results obtained by other authors who have used various other methods to solve the same problems.
Cristina Laura Sirbu, Politehnica University of Timisoara
Time: 4:00 pm - 4:20 pm (CST)
Title: Human Emotional State Detection
Abstract:
Emotional states affect human actions or behavior. Detecting them may be important in some situations. One may prefer not to approach an angry person. On the other hand, people may want to comfort sad persons. In automotive, driving under emotional states may lead to accidents. If the Driver Monitoring System detects an angry driver, it may limit the speed.
Humans have no problem recognizing emotional status of others, just by looking at pictures or videos. This ability can be extended to technical systems. This paper presents a method to detect the anger, the joy, the sadness, the surprise, and neutral state based on the facial elements.
Humans are pretty good at detecting emotions from facial elements (these are eyebrows, eyes, nose and mouth). The facial landmarks are detectable from images [1].
(Joint work with Ion Rareş STANCIU, Cătalin CĂLEANU)
Reference
[1] Kazemi, V., Sullivan, J., "One millisecond face alignment with an ensemble of regression trees", IEEE Conference on Computer Vision and Pattern Recognition, pp. 1867-1874, 2014.
Silviu Vert, Politehnica University of Timisoara
Time: 4:20 pm - 4:40 pm (CST)
Title: User Experience of Mobile Augmented Reality Applications in Cultural Heritage: a Case Study on Digital Storytelling with Historical Images in Urban Contexts
Abstract:
Augmented reality is a proper medium for preserving, documenting, and exploring all the values that cultural heritage holds. As a technology that reached mainstream audiences only in the recent years, augmented reality still lacks enough user experience research, in general but more specifically in the cultural heritage field, which is the context of this paper. Starting from tourist requirements and broad user experience models of such applications, as determined in the scientific literature, in this paper we describe a user experience framework for mobile augmented reality applications in cultural heritage, with a focus on digital storytelling with historical images in urban contexts. Our case study presents the process of development and validation of such a framework; the integration of mobile eye tracking in the usability evaluation toolkit for mobile augmented reality applications in cultural heritage; and the implementation of a mobile application with robust tracking of urban buildings for digital storytelling with historical images. The practical impact of our research is in improving the local application Spotlight Heritage Timisoara AR, that we developed as part of the Timisoara 2023 European Capital of Culture program in Romania.
(Joint work with Oana Rotaru, Ciprian Orhei, Victor Holotescu, and Diana Andone)