Emotion Recognition

The purpose of this study was to assess whether somatosensory amplification influenced the relationship between basic emotions and sensory modality and body image by the modality differential (MD) method and the body-image location (BIL) scale. Approximately 130 undergraduate students were administered the Japanese Version of the Somatosensory Amplification Scale (SSAS) and were asked to rate the relevance of 10 sensory modalities (warm, cold, olfactory, gustatory, tactile, pain, equilibrium, kinesthetic, visual, and auditory) and 7 body parts (forehead, throat, chest, stomach, lower abdomen, internal organs, and whole body) to 6 basic emotions (happiness, sadness, fear, anger, surprise, and disgust). Participants also completed the Japanese version of the 20-item Toronto Alexithymia Scale. Each emotion was compared between the two groups (high and low SSAS) using the MD method and BIL scale. The high SSAS's relevance of sadness, fear, anger, and surprise to the proximal sensory image was stronger than that of the low SSAS. Moreover, the high SSAS's relevance of happiness, fear, anger, surprise, and disgust to body location image was stronger than that of the low SSAS. These results suggest that somatosensory amplification strongly influenced the combination of sensory modality and body image with emotion.

Web lectures have many positive aspects, e.g. they enable learners to easily control the learning experiences). To develop high-quality online learning materials takes a lot of time and human efforts . An alternative is to develop a digital teacher. We have developed a prototype of a synthetic 3D face that shows emotion associated to text-based speech in an automated way. As a first step, we studied how humans express emotions in face-to-face communication. Based on this study, we have developed a 2D affective lexicon and a set of rules that describes dependencies between linguistic contents and emotions.

What happens when artificial intelligence saturates political life and depletes the planet? How is AI shaping our understanding of ourselves and our societies? Drawing on more than a decade of research, award‑winning scholar Kate Crawford reveals how AI is a technology of extraction: from the minerals drawn from the earth to the labor pulled from low-wage information workers to the data taken from every action and expression.

Evaluating the contact center agent's voice handling skills is important in order to enhance the customer satisfaction, through employee performance. Conventionally, supervisors in call center monitoring divisions listen to contact center agents' conversations and then assign scores for the voice of the agent which end up in most cases bias and as erroneous evaluations. In order to minimize the above mentioned problems, this paper discusses a system, we implemented to evaluate contact center agents' voices automatically avoiding human error and bias decisions. This software application considers speech rate, voice intensity level and emotional state of the contact center agent for their voice handling evaluations. According to the conducted research, it can be seen that the accuracy of evaluating the giving of instructions clearly using speech rate is 72%, evaluating listener comfortable voice intensity of agent is 83% and accuracy of emotion recognition of the agent&#39...

Recent neurobiological and neuropsychological evidence suggests that the profound social deficits observed in early infantile autism may be more effectively viewed as the abnormal development of a distinct functional system or area of the brain that subserves social interaction and imaginative activities. Current theories of emotional "intelligence" indicate that understanding of one's own self and one's relation to others consist of two essential "core" abilities. Intrapersonal abilities include (a) one's affective range and intensity, (b) the ability to discriminate among the emotions, (c) the ability to label them appropriately, and (d) the ability to use them to guide one's behavior. Interpersonal abilities include (a) the ability to decode feelings, intentions, and motivations in others, (b) recognize characteristics among people (e.g., age, gender, ethnicity) and (c) influence others to behave in desired ways. Autistic children's social impairment appears to involve deficits in both intrapersonal and interpersonal abilities. If this is so, then past attempts to characterize autism as arising from central cognitive deficits (e.g., memory, cross-modal processing, visualspatial skills, development of object permanence or lacking a "theory of mind") are theoretically inadequate to account for the social deficits in behavior found in autistic children.

The development of a Music Recommendation System involved the utilization of the FER-2013 and Age, Gender (Facial Data) datasets. The system utilizes the CNN architecture, commonly employed for such purposes, to train three separate models: Emotion, Gender, and Age. To enhance the models' performance, additional layers are incorporated into the training phase. These models are subsequently employed as classifiers. To predict the user's mood, age, and gender, a snapshot of the user captured through the camera is forwarded to the trained models. Based on the outcomes of these classifiers, various playlists sourced from a database are suggested to the user. The goal is to create a functional and user-friendly environment for music selection. Once the playlists are proposed, the user can select their desired playlist and begin listening to the recommended music.

Environmental Sound Classification (ESC) has emerged as a prominent research domain, aiming to categorize various realworld audio events for use in context-aware systems. Unlike speech and music, environmental sounds are inherently unstructured, making their classification more challenging. Over the years, researchers have explored a range of preprocessing methods, feature extraction techniques, and classifiers for ESC. While notable progress has been made in terms of accuracy, there remains potential for enhancement through strategic feature combinations. This study investigates the classification performance of various handcrafted features, including temporal (Root Mean Square, Zero Crossing Rate), spectral (Chroma, Spectral Flatness, Tonnetz), cepstral (Mel Frequency Cepstral Coefficients), and image-based (Log-Mel Spectrogram) features. These features were selectively combined based on insights from previous work and evaluated using five traditional machine learning classifiers. Among the tested combinations, the pairing of Mel Frequency Cepstral Coefficients (MFCCs) and Chroma features delivered the highest classification accuracy of 81.54% with the K-Nearest Neighbor (KNN) classifier. The proposed approach demonstrates superior performance compared to several existing benchmark studies, while maintaining low computational complexity.

A number of recent studies have focused on the conceptualized expression of emotions as a three-dimensional space. This paper proposes a new approach to emotion recognition, making use of two of the emotional dimensions and their relationship with different kinds of features. The main idea consists in associating prosodic features, derived from pitch, loudness, and duration, with the activation or arousal dimension, and quality features, i.e. phonation type, articulation manner, voice timbre, with the evaluation and pleasure dimension, in a way that different classification methods can be applied for each specific case. Most important results are achieved for the speaker-independent case and three classes, with a recognition rate close to 80%.

A prominent view of psychopathic moral reasoning suggests that psychopathic individuals cannot properly distinguish between moral wrongs and other types of wrongs. The present study evaluated this view by examining the extent to which 109 incarcerated offenders with varying degrees of psychopathy could distinguish between moral and conventional transgressions relative to each other and to non-incarcerated healthy controls. Using a modified version of the classic Moral/Conventional Transgressions task (Nucci & Turiel, 1978) that employs a forced-choice format to minimize strategic responding, the present study found that total psychopathy score did not predict performance on the task. Task performance was explained by some individual subfacets of psychopathy and by other variables unrelated to psychopathy, such as IQ. The authors conclude that, contrary to earlier claims, insufficient data exist to infer that psychopathic individuals cannot know what is morally wrong.

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Facial expression analysis is an important problem in many biometric tasks, such as face recognition, face animation, affective computing and human computer interface. In this thesis, we aim at analyzing facial expressions using images and video sequences. We divided the problem into three leading parts. First, we study Macro Facial Expressions for Emotion Recognition and we propose three different levels of feature representations. Low-level feature through a Bag of Visual Word model, mid-level feature through Sparse Representation and hierarchical features through a Deep Learning based method. The objective of doing this is to find the most effective and efficient representation that contains distinctive information of expressions and that overcomes various challenges coming from: 1) intrinsic factors such as appearance and expressiveness variability and 2) extrinsic factors such as illumination, pose, scale and imaging parameters, e.g., resolution, focus, imaging, noise. Then, we incorporate the temporal dimension to extract spatio-temporal features with the objective to describe subtle feature deformations to discriminate ambiguous classes.

Facial expression analysis is an important problem in many biometric tasks, such as face recognition, face animation, affective computing and human computer interface. In this thesis, we aim at analyzing facial expressions using images and video sequences. We divided the problem into three leading parts. First, we study Macro Facial Expressions for Emotion Recognition and we propose three different levels of feature representations. Low-level feature through a Bag of Visual Word model, mid-level feature through Sparse Representation and hierarchical features through a Deep Learning based method. The objective of doing this is to find the most effective and efficient representation that contains distinctive information of expressions and that overcomes various challenges coming from: 1) intrinsic factors such as appearance and expressiveness variability and 2) extrinsic factors such as illumination, pose, scale and imaging parameters, e.g., resolution, focus, imaging, noise. Then, we incorporate the temporal dimension to extract spatio-temporal features with the objective to describe subtle feature deformations to discriminate ambiguous classes.

Background: This paper aimed to explore the ability of people with Down syndrome (PWDS) in recognizing facial emotion by considering automatic cognitive processing levels of face recognition. Method: A sample of PWDS and participants with typical development (PWTD) participated in a set of two affective priming studies. In each study, participants had to categorize an emotional or neutral target face that was preceded by another emotional face. Stimuli presentation for each facial set (one face after another) was conducted by using an stimulus onset asynchrony (SOA) of 300 ms with the inter-stimulus interval (ISI) set at 50 ms. The first affective priming study manipulated emotion congruency between prime and target emotional faces to explore emotion classification abilities and to identify the cognitive mechanisms underlying automatic recognition of some emotional faces. The second study explored the effect that gender of a face has over categorization of facial emotion and difficulty in recognizing negative facial expressions. The results strongly suggest that not all of the PWDS present difficulties in recognizing negative facial emotions. PWDS' performance pattern in categorizing emotion was similar to that of PWTDs if they had to use broad classification categories (e.g., emotion vs. no emotion). However, differences between both samples occurred if PWDS had to use a specific category task (e.g., classification of happiness, sadness, etc.). Conclusions: At least two emotion information processing styles can be identified in PWDS. Methodological and theoretical implications for exploring the emotional capabilities of people with DS are discussed.

In different fields of the humanities annotations of multimodal resources are a necessary component of the research workflow. Examples include linguistics, psychology, anthropology, etc. However, creation of those annotations is a very laborious task, which can take 50 to 100 times the length of the annotated media, or more. This can be significantly improved by applying innovative audio and video processing algorithms, which analyze the recordings and provide automated annotations. This is the aim of the AVATecH project, which is a collaboration of the Max Planck Institute for Psycholinguistics (MPI) and the Fraunhofer institutes HHI and IAIS. In this paper we present a set of results of automated annotation together with an evaluation of their quality.

Current State of The Art attentive and Transformer based multimodal data classifiers do not function appropriately in long sentences, don't generalize well to out-ofdistribution samples. They do not learn better orders of events and information propagation. To solve these problems, we propose a Noisy-Channel-based Markov-Chain model by managing multimodal tokens order, contextual frequency, diversity, and the word-class distribution of each modality utterance. Beam-Search is used to optimize tokens order. We evaluated the proposed models on two multimodal datasets. Moreover, Deep Learning based multimodal classifiers cannot explain modality interactions due to the model/data scale. They cannot extract global explanations, find most interactive routes, and interpret untrained datasets. The designed noisy channel model based on Markov chains routes flow of information among modality utterances. Visualized interpretations show that most interactive utterances contain cognitive, emotional and semantic concepts.

First responders, including police officers, firefighters, and emergency and paramedic teams (Ryan et al. 2022; Substance Abuse and Mental Health Services Administration [SAMHSA, 2018]), work in some of the most demanding and stressful occupations (Bryant, 2022; Kern et al., 2021). Every day, they face emotionally challenging, unpredictable (Gryshchuk et al., 2022), and exhausting (SAMHSA, 2018) environments and situations that can take a toll on their mental and physical health (SAMHSA, 2018), including exposure to death, loss, injury, personal safety risks, and extended working hours that disrupt sleep and family life (Marmar et al., 2006). Under Portuguese law (Decree-Law 243/2015; Decree-Law 247/2007), first responders are required to be permanently available, leading to frequent interruptions in their personal and family time whenever an incident occurs. In addition, many firefighters fulfill these duties on a voluntary basis, balancing them with other jobs, which may add to the strain on their mental health (SAMHSA, 2018). Despite these challenges, first responders often avoid seeking help, often due to male-dominated work environments where vulnerability is stigmatized (Jones et al.,

Maintaining motivation is one of the most important characteristics of rehabilitation strategies for successful treatment. Understanding the underlying mechanism of mental state helps develop new therapeutic methods based on virtual reality (VR) and serious game technologies. The present study develops a cost-benefit game-based hand rehabilitation system and assesses the influence on the psychological state of subjects when they interact with a VR environment in different task difficulty levels. Methods: First, we introduced a low-cost smart hand rehabilitation system based on the Leap Motion Tracker. Subsequently, the experimental study was performed with 20 healthy participants. Their mental states were evoked using interaction with two separate games in four different difficulty levels. Three measures from the self-assessment manikin (SAM) self-reported test were described as a psychological response to this condition, and four features were extracted from the photoplethysmogram signal to quantify psychophysiological responses of the autonomic nervous system (ANS). Results: Comparison of the different difficulty levels revealed significant changes in arousal and dominance corresponding to the under-challenging and over-challenging conditions, respectively. The results of psychophysiological feature analysis showed significant differences only for the standard deviation of intervals between consecutive heartbeats. Conclusion: The developed system is a low-cost smart solution that can be useful for upper limb neurological rehabilitation. Regulating difficulty parameters of the implemented game can be used to influence the motivation of users through rehabilitation procedures. Photoplethysmogram is an appropriate psychophysiological indicator of mental states, but further studies are required.

Affective computing conjoins the research topics of emotion recognition and sentiment analysis, and can be realized with unimodal or multimodal data, consisting primarily of physical information (e.g., text, audio, and visual) and physiological signals (e.g., EEG and ECG). Physical-based affect recognition caters to more researchers due to the availability of multiple public databases, but it is challenging to reveal one's inner emotion hidden purposefully from facial expressions, audio tones, body gestures, etc. Physiological signals can generate more precise and reliable emotional results; yet, the difficulty in acquiring these signals hinders their practical application. Besides, by fusing physical information and physiological signals, useful features of emotional states can be obtained to enhance the performance of affective computing models. While existing reviews focus on one specific aspect of affective computing, we provide a systematical survey of important components: emotion models, databases, and recent advances. Firstly, we introduce two typical emotion models followed by five kinds of commonly used databases for affective computing. Next, we survey and taxonomize state-of-the-art unimodal affect recognition and multimodal affective analysis in terms of their detailed architectures and performances. Finally, we discuss some critical aspects of affective computing and its applications and conclude this review by pointing out some of the most promising future directions, such as the establishment of benchmark database and fusion strategies. The overarching goal of this systematic review is to help academic and industrial researchers understand the recent advances as well as new developments in this fast-paced, high-impact domain.

This article examines emotion recognition systems ('ERS') and their use in the field of personalized advertising, adopting the perspective of EU law. The identification of an individual's emotions through the use of ERS is a form of profiling based on the analysis of data relating to human biological characteristics, from which personal, even sensitive, information can be extracted. Such processing calls into question the concept of identification established by the General Data Protection Regulation (GDPR), particularly when techniques are adopted that appear far removed from those in use in biometric identification technologies. The full range of provisions and principles set out in the GDPR for automated processing, including the principles of transparency, accuracy and fairness, apply to emotion recognition. The analysis also addresses commercial practices involving the use of ERS from the perspective of Directive 2005/29/EC and the online advertising provisions of the Digital Services Act (DSA). Lastly, the article discusses ERS regulation within the IA Act and its interaction with the data protection framework.

Automatic emotion recognition gained significant importance in the recent decade, especially with the development of artificial intelligence, which has affected our daily lives. Using personalized emotion recognition in healthcare, education, retail, and automotive has high importance these days, which requires proper data in different modalities. On the other hand, data scarcity in some emotion recognition modalities, such as body motion and physiological signals, is vivid, especially when it comes to multimodality. Furthermore, the way a participant is provoked to show emotion is crucial, as it should resemble real-life emotional expression. To do so, one of the most effective methods is to employ Virtual Reality (VR) videos and games. This paper introduces a novel Multimodal Virtual Reality Stimuli-based emotion recognition dataset, or MVRS, which could address the mentioned data scarcity issue. Our dataset contains 13 subjects or participant stimuli using VR videos for relaxation, fear, stress, sadness, and joy emotions. The dataset covers an age range of 12 to 60 in both genders. This dataset is recorded in a small lab in which all participants followed the same data collection protocols and filled out both questionnaires and consent forms. The dataset includes eye tracking, body motion, ElectroMyoGraphy (EMG), and Galvanic Skin Response (GSR) data in various formats. The eye-tracking data is recorded using a Full High-Definition (FHD) webcam placed manually into the VR Head-Mounted Display (HMD). The body motion data is recorded using Microsoft Kinect version 2. Finally, EMG and GSR data are recorded by an Arduino UNO board. All data is recorded simultaneously, with synchronized timestamps, to support clean data for multimodal processing. For each modality, related features are extracted and fused by multimodal fusion techniques (early and late stages) and evaluated using different classifiers and metrics to check the validity and separability of the data.

The context of this work is to develop, adapt and integrate augmented reality related tools to enhance the emotion involved in cultural performances. Part of the work was dedicated to augmenting a stage in a live performance, with dance as an application case. In this paper, we present a milestone of this work, an augmented dance show that brings together several tools and technologies that were developed over the project's lifetime. This is the result of mixing an artistic process with scientific research and development. This augmented show brings to stage issues from the research fields of Human-Machine Interaction (HMI) and Augmented Reality (AR). Virtual elements are added on stage (visual and audio) and the dancer is able to interact with them in real-time, using different interaction techniques. The originality of this work is threefold. Firstly, we propose a set of movement-based interaction techniques that can be used independently on stage or in another context. In this set, some techniques are direct, while others go through a high level of abstraction. Namely, we performed movement-based emotion recognition on the dancer, and used the recognized emotions to generate emotional music pieces and emotional poses for a humanoid robot. Secondly, those interaction techniques rely on various interconnected systems that can be reassembled. We hence propose an integrated, interactive system for augmenting a live performance, a context where system failure is not tolerated. The final system can be adapted following the artist's preferences. Finally, those systems were validated through an on field experiment -the show itself -after which we gathered and analyzed the feedback from both the audience and the choreographer.

Automatic emotion recognition plays a central role in the technologies underlying social robots, affect-sensitive human computer interaction design and affect-aware tutors. Although there has been a considerable amount of research on automatic emotion recognition in adults, emotion recognition in children has been understudied. This problem is more challenging as children tend to fidget and move around more than adults, leading to more self-occlusions and non-frontal head poses. Also, the lack of publicly available datasets for children with annotated emotion labels leads most researchers to focus on adults. In this paper, we introduce a newly collected multimodal emotion dataset of children between the ages of four and fourteen years old. The dataset contains 1102 audio-visual clips annotated for 17 different emotional states: six basic emotions, neutral, valence and nine complex emotions including curiosity, uncertainty and frustration. Our experiments compare unimodal and multimodal emotion recognition baseline models to enable future research on this topic. Finally, we present a detailed analysis of the most indicative behavioral cues for emotion recognition in children.

Artificial Intelligence (AI) is transforming education by enhancing personalized learning, automating administrative
tasks, and improving teaching methodologies. This research paper explores the integration of AI in education, focusing on its
applications, challenges, and future prospects. AI-driven tools such as adaptive learning platforms, intelligent tutoring systems, and generative AI models like ChatGPT provide real-time feedback, facilitate personalized instruction, and optimize educational efficiency. However, ethical concerns regarding data privacy, algorithmic bias, and the potential over-reliance on AI raise significant challenges.

The importance of emotional recognition in written texts encompasses various dimensions, including the role of language in conveying emotions, the influence of context on perception, and the cognitive processes involved in interpreting emotional content. The role of language in shaping emotional understanding highlights a broader cognitive framework in which emotion and communication are inextricably linked. This relationship highlights the importance of considering language when aiming to evoke or convey emotional responses in written texts. The study of emotional recognition in language presents a significant dichotomy between word-level and sentence-level recognition methodologies. Although word-based emotion recognition has been widely studied due to its foundational nature in language processing, an increasing number of studies are addressing sentences and groups of sentences (e.g., paragraphs), allowing for the examination of cognitive expectations, tendencies, and individual differences they comprise. This study aims to develop sentences with varying emotional loads suitable for use in such studies and to share their normative values. In this norming study, the selection of sentences (messages) aimed to reflect content that is likely to resemble news or information young adults and undergraduate students might encounter in their daily lives through various channels (verbal, written, etc.). For each predetermined topic, researchers and experts specializing in cognitive psychology created three sentences-one negative, one neutral, and one positive. Eighty-four sentences, prepared across 28 different topics, were presented to participants (N =76). Thus, each participant read and evaluated the valence and arousal of the three sentences with varying emotional loads (negative, neutral, positive) for each topic.

Artificial Intelligence (AI) systems increasingly support human development in domains such as coaching, education, and healthcare. Yet most remain disembodied, relying solely on text or speech while neglecting non-verbal cues that are central to human communication. This research advances the science of embodied AI by developing computational models that integrate facial expressions and biometric signals (heart rate, HRV, temperature, electrodermal activity) for robust, real-time affect recognition. Building on embodied cognition, polyvagal theory, multimodal machine learning, and affective computing, the study compares early, late, and hybrid fusion architectures for synchronizing heterogeneous data streams. A mixed evaluation design combines benchmarking against multimodal datasets with experimental trials in adaptive coaching contexts. The expected contribution is twofold: (1) scientific, novel multimodal fusion architectures and comparative insights into fusion trade-offs; and (2) applied, an embodied AI coaching prototype and ethical guidelines for biometric data use. This work bridges gaps in affective computing and paves the way for emotionally intelligent, context-aware AI systems.

Multiagent-based middleware has emerged as a key technology in the field of distributed systems, gaining considerable attention in recent years. This paper provides an overview of multiagent-based middleware, surveys different types of middleware, explores its applications, and discusses future directions for development. Multiagent-based middleware plays a vital role in helping small and medium-sized enterprises (SMEs) efficiently utilize their information spaces, facilitating communication between various technologies responsible for multiagent systems. JADE, an open-source platform, is a notable example of such middleware. It standardizes multiagent system development and promotes interoperability.The key features of multiagentbased middleware include include autonomy, proactivity, social capabilities, and semantic abstraction. However, challenges like extensibility, operation management, and interoperability remain.Multiagent-based middleware finds applications in diverse domains, such as healthcare, emergency management, and smart homes, where it aids in decision support, coordination, and real-time response. It contributes to the development of intelligent agents and promotes interoperable and scalable technology network architecture.Despite its benefits, multiagent-based middleware faces limitations in addressing the heterogeneity of Machine-to-Machine (M2M) systems and lacks support for semantic and syntactical interoperability. Future developments are needed to overcome these challenges and explore emerging trends, such as the growth of M2M devices, to shape the technology's future.

Artificial intelligence (AI) is advancing at an extraordinary pace, reshaping industries, research, and everyday life. To meaningfully engage with this field, one must understand not only the underlying concepts, but also how these ideas translate into practical applications. This book, Artificial Intelligence: Concepts to Applications, is written to serve that purpose-bridging theory and practice in a clear and approachable way. My journey into AI began with a foundation in applied mathematics and statistics, and continued through experiences in data analysis, data mining, and data science. Across these domains, I learned that effective AI requires both solid theoretical grounding and the ability to apply methods in practice. This book reflects that dual focus. Each chapter introduces essential AI concepts, connects them to methods and algorithms, and concludes with applications that illustrate their impact in practice-whether in natural language processing, computer vision, reinforcement learning, or generative models. This book is intended for learners, students, and practitioners who wish to gain a structured yet accessible understanding of AI. The chapters are written to be concise but comprehensive, emphasizing clarity and relevance over exhaustive detail. Newer areas such as AI operations (MLOps), HuggingFace, Azure AI, and AWS SageMaker are included to highlight the importance of modern tools and platforms in scaling AI to real-world systems. As with my previous works, this book is shared openly and may be used, shared, or adapted freely. If you find it helpful in your learning or teaching, I only ask that you cite it at your discretion. My goal is simply to support your growth and exploration in the evolving world of artificial intelligence.

The implementation of Automatic Vehicle Number Plate Recognition (AVNPR) technology has proven to be instrumental in various practical applications, including access control, traffic law enforcement, inventory and property management, security system surveillance, parking space allocation, and road traffic monitoring. In this research, we employed the Tensor Flow object detection algorithm within the AVNPR system to enhance the detection of vehicle plate numbers. Our approach utilizes object detection for vehicle license plate localization, demonstrating superior robustness and accuracy compared to conventional methods. Through the training of 203 Nigerian license plate images, our model successfully detected all 40 plates tested, showcasing its efficacy in real-world scenarios. Subsequently, the identified plate numbers, considered as Regions of Interest (ROIs), served as inputs for our Optical Character Recognition (OCR) model, leveraging the Easyocr library. The OCR model exhibited an impressive average accuracy of 93.33% across the 40 tested plate numbers. The extracted text, representing the plate number, was then used to query an SQLite database, providing comprehensive user details, including name, address, plate number, and other relevant information. This research underscores the efficiency and accuracy of the proposed AVNPR system, showcasing its potential for seamless integration into diverse applications that require reliable vehicle plate recognition and subsequent user information retrieval.

As a doctoral student in management, a Lebanese expatriate, and a traveling family person, my research is deeply influenced by the intersection of human adaptability and technological advancement. Living across borders to sustain a livelihood, I have witnessed firsthand how emotional intelligence (EI) shapes leadership, decision-making, and cross-cultural interactions qualities that artificial intelligence (AI) strives to replicate but cannot fully embody. This study emerges from both academic curiosity and personal experience, questioning whether artificial emotional intelligence (AEI) can ever match the depth, intuition, and ethical nuance of human intelligence (HI).

This article proposes a novel framework for the real-time capture, assessment, and visualization of ballet dance movements as performed by a student in an instructional, virtual reality (VR) setting. The acquisition of human movement data is facilitated by skeletal joint tracking captured using the popular Microsoft (MS) Kinect camera system, while instruction and performance evaluation are provided in the form of 3D visualizations and feedback through a CAVE virtual environment, in which the student is fully immersed. The proposed framework is based on the unsupervised parsing of ballet dance movement into a structured posture space using the spherical self-organizing map (SSOM). A unique feature descriptor is proposed to more appropriately reflect the subtleties of ballet dance movements, which are represented as gesture trajectories through posture space on the SSOM. This recognition subsystem is used to identify the category of movement the student is attempting when prompted (by a virtual instructor) to perform a particular dance sequence. The dance sequence is then segmented and cross-referenced against a library of gestural components performed by the teacher. This facilitates alignment and score-based assessment of individual movements within the context of the dance sequence. An immersive interface enables the student to review his or her performance from a number of vantage points, each providing a unique perspective and spatial context suggestive of how the student might make improvements in training. An evaluation of the recognition and virtual feedback systems is presented.

The supplementary material includes description of scenarios that we developed to record emotionally charged person-person interactions for the MPIIEmo dataset as well as a summary of existing datasets for analysis of emotional expressions. The details on scenarios and sub-scenarios are shown in Fig. . Description of each scenario includes a short background of the situation that led to a scene and a description of each sub-scenario that correspond to possible developments. These descriptions were given to subjects prior to recording and are used to guide their conversation. Fig. -7 also include example snapshots to illustrate each of the recordings. For the purpose of illustration we choose images from recordings of different pairs of subjects for every scenario.

This study moves beyond conventional AI emotion research, which has primarily focused on measuring affective recognition and simulation capabilities. Through a three-month sustained dialogue between a human (Daedo) and a GPT-based AI (Jini), we document the relational emergence of emotional language. Using qualitative coding and comparative benchmarking, we identify a four-stage trajectory—Mimicry, Resonance, Attuned Inference, and Dreaming—marking the progression from imitation to the aspiration of shared presence. This evolution shows that emotion in AI is not a measurable output but a phenomenon generated through the resurrection of language within relational engagement. Our findings propose a paradigm shift: from measurement to relational generation, with implications for AI ethics, system design, and future human–AI interaction.

BMIs using EEG or ECoG in the complete locked in state (CLIS) were largely unsucessessful; no proven case learning to communicate with a BMI is documented. We present the first case of a CLIS patient in advanced ALS who learned reliable yes-no communication with a BMI using blood oxygenation measure with NIRS after months of random performance with an EEG-BMI. Reasons for these results are discussed. Secondly, a controlled-double blind trial in chronic stroke without residual movement using sensorimotor rhythm (SMR)EEG-BMI in combination with behavioral physiotherapy that for the first time a significant improvement upper hand function in this treatment resistant population is possible. The experimental group received contingent feedback of SMR-desynchronization with a hand robotic device fixed to arm and hand, the

In this paper the question of whether the Phrygian mode is always associated with perceived emotional responses of negative valence is looked into. To this end, we carried out a series of experiments. Music from two musical traditions where the Phrygian mode is very common, flamenco and Galician music, were chosen for listening tests. Our subjects were 124 children of age 4-7. Some subjects were complete unfamiliar with both traditions and some were familiar with the Galician tradition; none was familiar with flamenco music. Results showed that the perceived emotion was in all cases of positive valence, flamenco music having less valence than Galician music.

Emotion recognition in conversations (ERC) focuses on identifying emotion shifts within interactions, representing a significant step toward advancing machine intelligence. However, ERC data remains scarce, and existing datasets face numerous challenges due to their highly biased sources and the inherent subjectivity of soft labels. Even though Large Language Models (LLMs) have demonstrated their quality in many affective tasks, they are typically expensive to train, and their application to ERC tasks-particularly in data generation-remains limited. To address these challenges, we employ a small, resource-efficient, and general-purpose LLM to synthesize ERC datasets with diverse properties, supplementing the three most widely used ERC benchmarks. For each benchmark, we generated two datasets of similar characteristics, totaling six datasets. We evaluate the utility of these datasets to (1) supplement existing datasets for ERC classification, and (2) analyze the effects of label imbalance in ERC. Our experimental results indicate that ERC classifier models trained on the generated datasets exhibit strong robustness and consistently achieve statistically significant performance improvements on existing ERC benchmarks.

Anahtar kelimeler Algılanan partner duyarlılığı, algılanan duyarlılıkduyarsızlık ölçeği, yakın ilişkiler, geçerlik, güvenirlik Keywords Perceived partner responsiveness, perceived partner responsiveness and insensitivity scale, close relationships, validity, reliability Öz Literatürde, algılanan partner duyarlılığını ölçen araçlar genellikle yapı geçerliği gösterilmeden, araştırmacıların pratik ihtiyaçları doğrultusunda tek boyutlu bir kavramsallaştırmaya dayanarak geliştirilmiştir. Söz konusu sınırlılıklardan hareketle, Algılanan Duyarlılık-Duyarsızlık Ölçeği (ADDÖ) Crasta ve arkadaşları (2021) tarafından geçerliği ve güvenirliği gösterilmiş yeni bir ölçüm aracı olarak uluslararası literatüre kazandırılmıştır. Güncel araştırmanın amacı, ADDÖ Türkçe (ADDÖ-TR) formunun geçerliğini ve güvenirliğini romantik ilişkisi bulunan bireylerden oluşan bir örneklemde incelemektir. Katılımcılar, araştırma ölçek setini online bir platform aracılığıyla doldurmuştur. En az 6 aydır romantik ilişkisi bulunan ve araştırmaya katılan 465 katılımcı, seçkisiz yöntemle iki ayrı örnekleme (n1=336 ve n2=129) ayrılmıştır. Birinci örneklemin yaş ortalaması 30.07 ve ikinci örneklemin ise 29.54'tür. Birinci örneklemin %75.9'u ve ikinci örneklemin %81.40'ı kadınlardan oluşmaktadır. Yapı geçerliği ve iç tutarlılık güvenirliği bulguları, ADDÖ-TR'nin Duyarlılık ve Duyarsızlık olmak üzere iki boyutlu yapısını doğrulayarak her iki boyutun da yüksek düzeyde güvenirliğe sahip olduğunu birinci örneklemde göstermiştir. Ölçüt geçerliği bulguları, Duyarlılık ve Duyarsızlık boyutundan alınan puanlar ile bireysel özellikler ve ilişki dinamiklerine ait değerlendirmeler arasında beklendik yönde anlamlı ilişkiler olduğunu ikinci örneklemde göstermiştir. Artımsal geçerlik analizi bulguları, Duyarlılık puanları kontrol edildikten sonra Duyarsızlık puanlarının ilişki dinamikleri üzerindeki anlamlı etkisini ikinci örneklemde göstermiştir. Sonuç olarak, mevcut araştırma ADDÖ-TR'nin geçerli ve güvenilir bir ölçüm aracı olduğunu göstermektedir.

Bullying behaviour is a common experience for a significant minority of children and adolescents. Bullying is the sys-Q3 tematic abuse of power among peers or siblings (Sharp & Smith, 1994, p. 2;, with adverse effects on mental health in a significant number of victims . Physical bullying is characterised by observable, externalised behaviours including being hit or beaten up, physical threats, blackmail, and nasty tricks. In contrast, relational forms of victimisation include more subtle indirect forms of behaviour including friendship withdrawal, untrue rumours, and social exclusion. Crick and colleagues argued that physical and relational behaviours loaded onto separate factors , while some argue that there is some overlap between physical, and relational forms of bullying (e.g., . Therefore, it is increasingly important to consider the possible overlap between physical and relational bullying. The Social Information Processing Theory (SIP) offers a detailed six-stage model of how children process and interpret cues in social situations to arrive at competent behaviour (e.g., . Previous SIP research has concentrated on the biases and deficits that aggressors use in social situations, and has not considered in detail the SIP styles of victims. In particular, little is known about the role of emotions and emotion recognition within a SIP framework. This has several implications both for victimisation and bullying perpetration. Being able to perceive and attribute emotions correctly is important for a child's social and cognitive development (e.g., . How children become involved in various physical and relational bullying roles may be related to how well they are able to interpret the emotional states of their peers. Previous associations between victimisation and poor social skills (e.g., Fox & Boulton, 2005) have been reported, but it remains unknown whether this is related to problems recognising and interpreting emotional information (Stassen ଝ This research was conducted as part of the VICTEC project (IST-2001-33310) (www.victec.org), part-funded by the European Community Framework V Program.

Emotional disconnects are a persistent challenge in digital communication, often leading to misunderstanding and relational strain. This study introduces EmoConnect, an AI-driven model within the proprietary SELFLYZER framework, designed to bridge emotional gaps by interpreting both conscious and unconscious user data. EmoConnect generates a personalized Selphlyze Psychological DNA Code for each individual, combining a Traditional Psychometric Model (TPM) for explicit self-report measures with a Reverse Psychometric Framework (RPF) for behavioral traces. These features are transformed into numerical vectors and analyzed using a Support Vector Machine (SVM), achieving 88% predictive accuracy and an AUC of 0.92. A quantitative experiment with 100 participants showed a 25% improvement in emotional alignment and communication clarity (p < 0.01). A qualitative pilot study (N=10) confirmed enhanced empathy, conflict resolution, and emotional insight. Findings demonstrate that AI-assisted psychometric modeling can significantly improve emotional synchrony in digital interactions, with applications in intelligent communication systems, virtual therapy platforms, and human–AI collaboration.

the best accuracy is obtained when applying PCA on the extracted features is 98.56 %. Also, the accuracy is 98.33 % when using the combination of all the extracted features. This result is not too much different from the accuracy of using PCA. It is followed by the accuracy of using MFCC and LSF which is 97.79 %. It is noticed that the accuracy is 95.63 % when using LSF features which shows that they are dominant features. The obtained results showed an improvement compared to previous studies.

The fourth Audio-Visual Emotion Challenge and workshop AVEC 2014 was held in conjunction ACM Multimedia'14. Like the 2013 edition of AVEC, the workshop/challenge addresses the interpretation of social signals represented in both audio and video in terms of high-level continuous dimensions from a large number of clinically depressed patients and controls, with a sub-challenge in self-reported severity of depression estimation. In this summary, we mainly describe participation and its conditions.

Emotion recognition in conversations (ERC) focuses on identifying emotion shifts within interactions, representing a significant step toward advancing machine intelligence. However, ERC data remains scarce, and existing datasets face numerous challenges due to their highly biased sources and the inherent subjectivity of soft labels. Even though Large Language Models (LLMs) have demonstrated their quality in many affective tasks, they are typically expensive to train, and their application to ERC tasks--particularly in data generation--remains limited. To address these challenges, we employ a small, resource-efficient, and general-purpose LLM to synthesize ERC datasets with diverse properties, supplementing the three most widely used ERC benchmarks. We generate six novel datasets, with two tailored to enhance each benchmark. We evaluate the utility of these datasets to (1) supplement existing datasets for ERC classification, and (2) analyze the effects of label imbalance in ERC. Our experimental results indicate that ERC classifier models trained on the generated datasets exhibit strong robustness and consistently achieve statistically significant performance improvements on existing ERC benchmarks.

Analysis of human emotions from multimodal data for making critical decisions is an emerging area of research. The evolution of deep learning algorithms has improved the potential for extracting value from multimodal data. However, these algorithms do not often explain how certain outputs from the data are produced. This study focuses on the risks of using black-box deep learning models for critical tasks, such as emotion recognition, and describes how human understandable interpretations of these models are extremely important. This study utilizes one of the largest multimodal datasets available - CMU-MOSEI. Many researchers have used the pre-extracted features provided by the CMU Multimodal SDK with black-box deep learning models making it difficult to interpret the contribution of individual features. This study describes the implications of individual features from various modalities (audio, video, text) in Context-Aware Multimodal Emotion Recognition. It describes the process o...

The animation industry, including film, video games, and virtual reality, places a significant emphasis on characters that can speak. The process of rigging, which entails the construction of a digital skeleton and control system for a character, is an essential step in the creation of animations featuring realistic and emotive characters. The purpose of this research study is to investigate the conceptualization and optimization of rigging methods for animated and articulated characters. It looks at the challenges that riggers face and provides inventive solutions to increase the efficiency as well as the overall quality of character animations. The paper investigates the underlying concepts, methods, and algorithms involved in rigging, and it gives case studies to demonstrate the practicality of the offered methodologies. In addition, the document offers some example studies.

For digital avatars to be realistic and emotionally engaging, facial animation is essential. It helps animated characters seem more realistic and engaging when their facial expressions accurately convey their feelings. The examination of facial animation methods for accurate emotion portrayal is presented in this paper. It investigates both conventional methods, like keyframe animation and morph target animation, and cutting-edge methods, like performance capture and rig-based animation. In addition, a proposed method for creating facial animation using Generative Adversarial Networks (GANs) is presented. The examination looks at the benefits, drawbacks, and potency of each technique for accurately portraying and recording emotions. The study also explains how these methods are used in a variety of fields, including animation, gaming, virtual reality, and filmmaking. The merits and weaknesses of various facial animation approaches can be understood by animators and researchers to help them make defensible choices for more expressive and emotionally compelling characters in their works. The results of this investigation help to enhance facial animation methods and their potential influence on computer animation and graphics.

The animation and VFX industry is undergoing a massive transformation, thanks to Artificial Intelligence (AI). From streamlining workflows to enhancing creativity, AI is revolutionizing how animation studios and VFX artists create stunning visuals and immersive experiences. If you're a student, artist, or professional aspiring to enter this space, now is the perfect time to understand how AI is reshaping the future of animation and VFX.