Top

2025 | Book

Read chapter Read first chapter

Advances in Mobile Computing and Multimedia Intelligence

22nd International Conference, MoMM 2024, Bratislava, Slovak Republic, December 2–4, 2024, Proceedings

Editors: Pari Delir Haghighi, Solomiia Fedushko, Gabriele Kotsis, Ismail Khalil

Publisher: Springer Nature Switzerland

Book Series : Lecture Notes in Computer Science

Part of: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Table of Contents

About this book

This book constitutes the refereed proceedings of the 22nd International Conference on Advances in Mobile Computing and Multimedia Intelligence, MoMM 2024, held in Bratislava, Slovak Republic, during December 2–4, 2024.

The 10 full papers and 8 short papers in this book were carefully reviewed and selected from 34 submissions. They were organized in topical sections as follows: wearable and sensor-based data for human performance and interaction; mobile user experience, motivation, and behavior; medical and cognitive health applications; image, video, and multimedia processing; software and system intelligence.

Frontmatter

Wearable and Sensor-Based Data for Human Performance and Interaction

Frontmatter

Investigating Choking Under Pressure in Dance Performance with Motion and Physiological Information Analysis

Abstract

This study is the initial phase in developing coping mechanisms for individuals who struggle with ‘choking under pressure’ due to stress. We focus on examining the motion and physiological information of dancers in high-pressure performance and low-pressure practice environments. Our primary objective is to identify key differences in stress responses between individuals who manage pressure well and those who do not. Through a comparative analysis of the motion and physiological information, we seek to identify the disparities between those who are adept at handling choking under pressure and those who are not. To simulate a high-pressure performance environment, a dance audition will be conducted and the physical and mental state of the auditionees during the evaluation process will be monitored and analyzed. The results indicated that auditionees who were unable to cope with choking under pressure showed significantly reduced physical movements compared to those who were able to handle it.

Shuhei Tsuchida, Ayumi Ohnishi, Kae Mukai, Ken Watanabe, Katsumi Watanabe, Tsutomu Terada, Masahiko Tsukamoto

A Method for Estimating the Force Applied on the Forearm Using PPG Sensors

Abstract

This study proposes estimating forearm pressure using a PPG sensor in a smartwatch or activity meter. The method involves creating a regression model that estimates the applied pressure by analyzing changes in the local maximum values of pulse waves before and after the pressure is applied. Experiments with five subjects evaluated the method. The individual evaluation model, using only the user’s data, had an average coefficient of determination of 0.57 and an MAE of 1.60 Kg. The total evaluation model, using all subjects’ data, had a coefficient of determination of 0.53 and an MAE of 1.97 Kg.

Ryo Watabe, Kazuya Murao

Good Vibes! Towards Phone-to-User Authentication Through Wristwatch Vibrations

Abstract

While mobile devices frequently require users to authenticate to prevent unauthorized access, mobile devices typically do not authenticate to their users. This leaves room for users to unwittingly interact with different mobile devices. We present GoodVibes authentication, a variant of mobile device-to-user authentication, where the user’s phone authenticates to the user through their wristwatch vibrating in their pre-selected authentication vibration pattern. We implement GoodVibes authentication as an Android prototype, evaluate different authentication scenarios with 30 participants, and find users to be able to well recognize and distinguish their authentication vibration pattern from different patters, from unrelated vibrations, and from the pattern being absent.

Jakob Dittrich, Rainhard Dieter Findling

A Method for Embedding Information Into Acceleration Data Using Resonant Frequency Sound to Capacitive Accelerometers

Abstract

This paper presents a method to embed arbitrary bit sequences into acceleration data by attacking capacitive accelerometers with sound. Using a loudspeaker, sound data representing bit sequences is irradiated to the device, manipulating the acceleration data instantly and continuously. Evaluation experiments showed that the extraction accuracy of 10-bit binary information is 100% from multifunctional sensors placed on the table, 99.7% from a smartphone placed on the table, 100% from multifunctional sensors worn at the wrist, and 98% from a smartphone held in the hand.

Takeru Yokoyama, Kazuya Murao

Mobile User Experience, Motivation, and Behavior

Frontmatter

MEUSec – Method for Enhancing User Experience and Information Security

Abstract

Digital identity wallets enable the management of digital identities and verification documents such as ID cards and driving licences. This data can be stored efficiently in one place on user devices. Research shows that some of the existing digital identity wallets have user experience and information security deficits. Users struggle to understand the concept of digital identity wallets, personal information is often inadequately secured or released to untrusted parties. Moreover, user experience and information security might influence each other negatively. Hence, it is necessary to consider user experience and information security simultaneously, and to evaluate and improve them together. However, existing methods focus on either aspect and do not consider their interplay. In this paper, we present the MEUSec method to facilitate an analysis and improvement of user experience and information security of digital identity wallets.

Max Sauer, Christoph Becker, Andreas Oberweis, Simon Pfeifer, Jan Sürmeli

Correlation Between Gamification and Intrinsic Motivation with a Mobile Job-Market Application

Abstract

In this paper we study the influence of the degree a mobile app is gamified on the motivation of its users. In particular, we created three versions of the same job application app, one without gamification, one with modest gamification, and one with heavy gamification. We then asked 21 student participants about their motivation to use either of these versions. Statistical analysis shows that gamification indeed increases motivation, but there is no significant difference between modest and heavy gamification.

Niklas Grossmann, Helmut Hlavacs

Query by Trash: Encouraging Green Attitudes and Behavior Through Eco-News Retrieval in Smart Trash Bins

Abstract

This paper proposes a smart trash bin to promote pro-environmental attitudes and behavior. The proposed system automatically searches for and displays news articles that are relevant to the type of trash discarded. The device ranks news articles in consideration of both the discarded trash type and its potential to evoke a sense of environmental urgency or guilt. To achieve this ranking, we constructed a dataset of news articles annotated for their “crisis/guilt” sentiment and trained a ranking model using a learning-to-rank algorithm. The highest-ranked article is displayed on a screen installed in the device.

The results of a user study demonstrated that a prototype device presented an opportunity for participants to reflect on environmental issues and demonstrated a modest influence on the formation of pro-environmental attitudes.

Momo Takeuchi, Yoshiyuki Shoji, Yusuke Yamamoto

Evaluating the Impact of Color and Sound Combinations on Cognitive Performance in Virtual Reality

Abstract

Virtual Reality (VR) has expanded into various fields. This research explores the effects of virtual environments on task performance, with a focus on the combination of environmental elements such as colors and sounds and cognitive tasks. Participants completed tasks in virtual environments with varied auditory and visual stimuli. Findings show natural sound-color combinations enhance concentration and efficiency. This study emphasizes how sensory stimuli in VR can optimize learning and work environments, demonstrating VR’s potential as a transformative tool for society.

Ryoma Nakao, Tatsuo Nakajima

Medical and Cognitive Health Applications

Frontmatter

Mild Cognitive Impairment Prediction Using Facial and Speech Data

Abstract

Mild cognitive impairment (MCI) represents a transitional stage between the cognitive decline associated with normal aging and more severe conditions such as dementia. Early diagnosis of MCI is crucial for effective healthcare intervention. However, current detection methods are often costly and time-consuming. This study introduces a multimodal fusion network (MFN) designed to predict MCI more efficiently. The proposed network utilizes dual-stream ResNets to process both facial and speech features. These features, extracted from the convolutional and subsampling layers of the ResNets, are subsequently fused in a fully connected layer to generate the final prediction. The dataset comprises a total of 52 participant videos, with an equal distribution: 26 videos from participants with normal cognitive function and 26 videos from participants diagnosed with MCI. Experimental results demonstrate the effectiveness of this approach, with an F1 score of 0.89 across test participants.

Chien-Cheng Lee, Wei-Chieh Huang, Yi-Fang Chuang

Comparing Training of Sparse to Classic Neural Networks for Binary Classification in Medical Data

Abstract

Sparse Neural Networks are increasing in popularity and provide the opportunity for compact and efficient models for resource-constrained environments which are expanding as the number of IoT devices is increasing and as the Edge Computing and Fog paradigms are gaining traction. We investigate and evaluate sparsifying the training of Convolutional Neural Networks for the task of binary classification on medical datasets. We considered low (i.e., \(28\times 28\)) grey-scale resolution images that are memory-friendly and suitable for storing and analysing on lightweight devices. We found out that high sparsification strategies (above 75%) can achieve comparable performances with that of the fully connected counterpart while allowing for a reduction in inference time and peak memory usage, beneficial for resource-constrained environments part of Edge Computing. It is important to note that, as might be expected, after 90% sparsity, the performance can oscillate, and the results can vary significantly.

Laura Erhan, Antonio Liotta, Lucia Cavallaro

A Genetic Algorithm-Based Scheduling Method Considering Working Hours for Medical Doctors

Abstract

With the diversity of work styles in recent years, we need to solve issues such as the aging of the working-age population and the increasing responsibilities of caregiving. In particular, the medical field requires efficient and appropriate scheduling due to the lengthening of working hours caused by the shortage of human resources. Many researchers have addressed the Nurse Scheduling Problem (NSP). However, the scheduling problem for medical doctors is more difficult than NSP because they have more varied work arrangements and more stringent constraints than those of the NSP. In this paper, we propose a method to automatically generate work scheduling that considers the work hours of medical doctors. The proposed method classifies medical doctors into four types of work arrangements (morning shift, afternoon shift, semi-night shift, and night shift) and constructs rules to generate constraints for each work arrangement. In addition, the proposed method uses a genetic algorithm to generate the optimal work schedules for multiple medical doctors considering computer resources in heuristic search. The evaluation results showed that the proposed method can generate work schedules that satisfy as many contraints as possible.

Subaru Narahashi, Eiji Hirakawa, Akira Uchiyama, Yusuke Gotoh

Image, Video, and Multimedia Processing

Frontmatter

Application of Benford’s Law to the Identification of Non-authentic Digital Images

Abstract

This study evaluated Benford’s law for detecting non-authentic digital images by analyzing the first digits of pixel values after a discrete cosine transform (DCT). We analyzed 137 pairs of authentic and modified JPEGs using ROC curves, k-means clustering, chi-squared tests, and PCA. The results showed AUC values near 0.5, indicating low classification performance. The k-means algorithm had 49% precision with low completeness, and PCA revealed a significant overlap between the authentic and manipulated images. These findings suggest the limited effectiveness of Benford’s law alone, highlighting the need to integrate advanced image-processing methods and explore additional pixel-distribution features for the effective detection of non-authentic images.

Jaroslaw Kobiela, Piotr Dzierwa

Efficient Moving Object Detection from Ultra-High Resolution Omnidirectional Video

Abstract

With advancements in panoramic camera technology, resolutions have significantly improved [1], capturing distant objects more clearly. The Insta360 Titan, for instance, supports up to 11k resolution (\(10560\times 5280\)), offering unprecedented detail. However, current object recognition methods struggle with such ultra-high-resolution footage. This paper presents a novel approach for detecting moving objects in 11k panoramic videos from the Insta360 Titan. Our method involves downsampling and background subtraction to detect moving objects quickly and accurately. These regions are then cropped into smaller images for high-precision detection, reducing computational load. This technique, inspired by proxy methods in video editing, maintains result quality while easing processing burdens. Experiments demonstrate our method’s ability to accurately detect humans at distances up to 60 m, achieving 15 fps, thus proving its effectiveness for ultra-high-resolution panoramic footage.

Takuro Ohashi, Shohei Yokoyama

Evaluation of the Clustering Method Used to Analyze the Proximity of Mobile Devices Using Indirect Geolocation Indicators

Abstract

This study introduces a new methodology for clustering mobile devices to determine their geographical proximity without using direct geolocation data. Using the DBSCAN algorithm, we aim to identify significant spatial patterns while preserving user privacy. The dataset used includes information on device manufacturers and network interactions. Clustering was performed after data preprocessing, with DBSCAN effectively grouping devices in proximity. Our results show that this method successfully identifies clusters that reflect meaningful geographical relationships among the devices. The evaluation included calculating the average and maximum distances within the clusters, demonstrating the robustness of the method. Despite its effectiveness, the success of the method depends on the quality and completeness of the input data. Future research should explore additional data sources and refine used algorithms to enhance accuracy and efficiency.

Jaroslaw Kobiela, Piotr Urbaniec

Software and System Intelligence

Frontmatter

Cross-Project Software Defect Prediction Using Ensemble Model with Individual Data Balancing and Feature Selection

Abstract

The quality of software significantly influences its safety and security. With the rapid expansion of software development, the issue of coding quality has become increasingly critical. The manual and resource-intensive nature of error detection in software and its inherent unreliability underscores the necessity for automation. Consequently, a burgeoning interest is employing machine learning methods for software defect prediction. This study introduces a novel stacking software cross-project defect prediction model. Each weak classifier undergoes a learning process incorporating individual data balancing and feature selection techniques. The efficacy of the model was evaluated using accuracy and F1 score metrics on multiple project datasets sourced from the PROMISE repository. The application of the proposed model yielded a classification accuracy of 0.839 and an F1 score of 0.909, surpassing the average performance of single classifiers.

Vitaliy Yakovyna, Oleh Nesterchuk

AUTO-DataGenCARS+: An Advanced User-Oriented Tool to Generate Data for the Evaluation of Recommender Systems

Abstract

Context-Aware Recommender Systems (CARS) offer context-based suggestions that are particularly crucial in the tourism domain, where personalized experiences significantly enhance user satisfaction. However, the evaluation of CARS is a challenge, partly due to the scarce availability of appropriate datasets that fulfill a variety of evaluation purposes. For example, to evaluate CARS, we need datasets that incorporate context data, but in practice existing datasets provide very little contextual information.

This paper presents AUTO-DataGenCARS+, a graphical user-oriented tool designed to generate synthetic data for evaluating both Recommender Systems and CARS. Some of the relevant features of the tool include: a flexible definition of user profiles, user, item and context schemas; a realistic generation of ratings and item attributes; the possibility to mix real and synthetic datasets; functionalities for analyzing and evaluating existing datasets; and an extendable architecture for advanced users. We illustrate the benefits of AUTO-DataGenCARS+ through several examples and experimental evaluations.

María del Carmen Rodríguez-Hernández, Sergio Ilarri, Marcos Caballero, Raquel Trillo-Lado, Ramón Hermoso, Rafael del-Hoyo-Alonso

A Method for Eliminating False Positives of Acceleration-Based Gesture Recognition Using Eye Tracking

Abstract

This paper sets up a scenario where a monitor is operated through gestures and proposes a method that excludes false positives in gesture recognition using eye tracking data. A preliminary experiment confirmed that existing gesture recognition using acceleration data often misrecognizes relatively simple gestures like Circle and Check during activities involving significant movements, such as stretching and walking, compared with more complex gestures like Cross and Triangle. Additionally, eye tracking data were collected during daily activities. Analyzing these results, we hypothesized that a user’s gaze movements are small during intentional gestures aimed at operating a monitor. By calculating the standard deviation of various gaze features and excluding segments with abnormal values, we aimed to enhance the accuracy of gesture recognition. The results showed that features such as Fixation Movement Distance and Average Pupil Diameter Change are relatively effective in excluding false positives, particularly for simple gestures like Circle and Check.

Hinase Kawano, Kazuya Murao

Toward the Implementation of a Cooking Support System Complementing Nonexistent Objects with Virtual Objects

Abstract

Cooking requires the usage of various ingredients and cooking utensils. However, if we lack some of them due to insufficient preparation, we cannot cook as expected. In this paper, we propose a cooking support system that complements nonexistent objects with virtual objects. We developed a prototype for cutting apples to evaluate the differences between using the real object and the virtual one. The results showed that the cutting technique equally enhanced both the cases of using a real knife and a virtual knife.

Taiki Nihanda, Shoji Sano

Backmatter

Title: Advances in Mobile Computing and Multimedia Intelligence
Editors: Pari Delir Haghighi
Solomiia Fedushko
Gabriele Kotsis
Ismail Khalil
Publisher: Springer Nature Switzerland
Electronic ISBN: 978-3-031-78049-3
Print ISBN: 978-3-031-78048-6
DOI: https://doi.org/10.1007/978-3-031-78049-3