Sentiment analysis from video streams requires effective modeling of both temporal dynamics and multimodal cues. This paper presents a lightweight multimodal CNN-LSTM framework, designed for low-latency operation on live video streams, that classifies individuals' affective states as negative, neutral, or positive by jointly exploiting audio and visual information. The proposed system consists of four lightweight deep neural network models that fuse sequences of audio and visual features into a unified temporal stream, enabling the accumulation of sentiment evidence over time for robust prediction. To further enhance discriminative power, dynamic mouth state classification is incorporated, allowing the model to better capture subtle visual cues related to speech and sentiment. Experiments conducted on the Ryerson Audio-Visual Database of Emotional Speech (RAVDESS) demonstrate that the proposed approach achieves a test accuracy of 84.89% on the three-class sentiment recognition task, competitive with substantially larger audio-visual architectures, while using an order of magnitude fewer parameters and maintaining low-latency operation. In addition, we provide a comprehensive comparison with multiple multimodal large language models (LLMs), evaluated under diverse input configurations and prompting strategies. The results offer practical insights into the strengths and limitations of contemporary LLMs for audio-visual sentiment analysis and highlight the continued effectiveness of compact, task-specific neural architectures for streaming audio-visual input.
Introduction Vehicular driving pattern analysis is important for improving road safety, driving efficiency, and intelligent transportation systems. The aim of this study is to analyze recent research on the major driver-behavior analysis tasks, which are driving pattern recognition, distraction detection, abnormal driving-behavior detection, driver profiling and identification, fuel-efficiency analysis, and driving-risk assessment. Methods The literature selection process was structured and included identification, screening, and eligibility. One hundred seventeen studies were used in this study, and 12 studies were selected for a detailed examination for comparative analysis. Results The studies analyzed have shown that driver-behavior analysis is becoming more popular with the use of machine learning, deep learning, sensor fusion, edge computing, and other new technologies. Existing approaches, however, are still scattered across behavioral tasks and encounter issues such as real-time processing, privacy preservation, scalability, model generalization, and integration of heterogeneous data. Discussion This study highlights the need for integrated and task-oriented approaches to address these challenges and support multiple behavioral assessments. The findings provide research directions on the development of scalable privacy-aware real-time driver-behavior analysis systems for connected and intelligent transportation environments.
Introduction Deception elicits measurable cognitive and linguistic changes, yet accurately identifying deceptive language in high-stakes environments remains a persistent challenge. Although natural language processing (NLP) and machine learning (ML) techniques have shown promise in detecting misinformation and low-stakes deception, their performance on spontaneous, stress-induced speech remains poorly understood. This study investigates linguistic markers of deception and evaluates the performance of supervised ML models and a knowledge-grounded language model, FranklinAsk™, across both synthetic and real-world datasets. Methods Linguistic analyses were conducted on responses from a National Center for Credibility Assessment (NCCA) military training exercise involving truthful and deceptive participants. A synthetic dataset was generated using MOSTLYAI to provide a balanced corpus for ML training. Linguistic features were extracted using pretrained NLP models, and supervised classifiers, including logistic regression, Random Forest, Naïve Bayes, and Support Vector Machine (SVM), were trained using LIWC-22 features with 10-fold cross-validation. The FranklinAsk™ RAG system was evaluated using the same participant responses to compare knowledge-grounded reasoning with statistical learning. Results Two linguistic features, cardinal numbers (nominal p = 0.041) and particles (nominal p = 0.0486), reached the unadjusted significance threshold, each with a moderate effect size ( d = 0.55); neither remained statistically significant after Benjamini-Hochberg FDR correction (adjusted p = 0.668 for both). No significant sentiment differences were observed. On the synthetic dataset, logistic regression performed best (F1 = 0.930; ROC AUC = 0.978), but all supervised models showed substantial performance declines when applied to the real-world dataset. In contrast, FranklinAsk™ achieved the highest F1-score (0.519), highest accuracy (0.487), and second-highest recall (0.700) on the military exercise data. Discussion These findings highlight the limited generalizability of synthetic-trained deception models to real-world, high-stakes communication and motivate further investigation of knowledge-grounded language models as an alternative approach when labeled real-world training data are limited. Future research should prioritize expanding ecologically valid datasets and exploring hybrid approaches that combine statistical modeling with knowledge-grounded reasoning.
