The internet of Things (IoT) technology's swift advancements have contributed to Wi-Fi signals being widely used in the acquisition of trajectory signals. The methodology of indoor trajectory matching aims to observe and analyze the movements and encounters between individuals in indoor spaces, thereby enabling a more thorough monitoring system. Due to the restricted computational power of IoT devices, cloud computing is essential for indoor trajectory matching, yet this also raises privacy concerns. This paper, therefore, advances a trajectory-matching calculation method capable of supporting ciphertext operations. Hash algorithms and homomorphic encryption are utilized to secure various private data, while trajectory similarity is calculated based on correlation coefficient analysis. Original data acquisition in indoor environments might be affected by obstacles and other interferences, causing some parts of the data to be missing. This paper also supports the recovery of missing ciphertext values via the mean, linear regression, and KNN methodologies. Employing these algorithms, the missing segments of the ciphertext dataset are forecast, ultimately yielding a complemented dataset with an accuracy exceeding 97%. This paper introduces novel and improved datasets for matching calculations, illustrating their practical feasibility and effectiveness in real-world scenarios, specifically regarding calculation time and precision.
When using eye movements to operate an electric wheelchair, unintentional actions like surveying the surroundings or studying objects can be mistakenly registered as control commands. Recognizing visual intent is paramount, as this phenomenon is known as the Midas touch problem. A deep learning model for real-time visual intent estimation, coupled with a novel electric wheelchair control system, is presented in this paper, incorporating the gaze dwell time method. The model proposed here is a 1DCNN-LSTM, which calculates visual intention by leveraging feature vectors from ten variables such as eye movements, head movements, and distance to the fixation target. Experiments evaluating visual intentions, categorized into four types, demonstrate the proposed model's superior accuracy compared to alternative models. The proposed model, applied to the electric wheelchair's driving tests, reveals a diminished user operating burden and an improvement in the wheelchair's manageability, when measured against the conventional method. From the collected data, we inferred that the learning of time-based patterns in eye and head movements can facilitate a more accurate prediction of visual intentions.
The advancement of technologies in underwater navigation and communication, while promising, does not readily overcome the difficulty in determining precise time delays for signals travelling substantial distances underwater. The paper introduces a refined method to quantify time delays with high accuracy in lengthy underwater sound propagation paths. Encoded signals initiate the signal acquisition process at the receiving station. At the receiving end, bandpass filtering is employed to enhance the signal-to-noise ratio (SNR). In light of the unpredictable variations in the underwater acoustic channel, a technique for selecting the optimal time window for cross-correlation is proposed. The calculation of cross-correlation outcomes is the subject of new proposed regulations. In order to ascertain the algorithm's effectiveness, we subjected it to a comparative analysis with other algorithms, leveraging Bellhop simulation data from low signal-to-noise ratio conditions. By the end of the procedure, the exact time delay was determined. Experiments conducted underwater at various distances support the high accuracy of the method suggested by the paper. The calculation deviates by approximately 10.3 seconds. The proposed method provides a contribution to the fields of underwater navigation and communication.
The demanding nature of modern information societies subjects individuals to persistent stress, a product of multifaceted work environments and intricate interpersonal relationships. The practice of aromatherapy, employing fragrant essences, is drawing considerable interest for its stress-alleviating properties. For a comprehensive understanding of aroma's influence on the human psychological state, a quantitative method of assessment is required. This study introduces a method for assessing human psychological states during aroma inhalation, employing two biological indices: electroencephalogram (EEG) and heart rate variability (HRV). The study's purpose is to analyze the interplay between biological indices and the psychological consequences of applying various scents. Seven different olfactory stimuli were used in an aroma presentation experiment, during which EEG and pulse sensor readings were captured. From the experimental data, we isolated and quantified EEG and HRV indexes, subsequently scrutinizing them in light of the olfactory stimuli presented. Our findings suggest that olfactory stimuli strongly affect psychological states during aroma stimulation. The human response to such stimuli is immediate, yet gradually becomes more neutral. The EEG and HRV indices exhibited significant differences in response to aromatic and disagreeable odors, specifically among male participants between 20 and 30 years of age. Meanwhile, the delta wave and RMSSD parameters indicated a potential for broader utilization of this method to evaluate psychological states affected by olfactory stimuli across different age groups and genders. XL765 Using EEG and HRV, the results indicate the potential for evaluating psychological responses triggered by olfactory stimuli like aromas. Additionally, an emotion map visualized the psychological states influenced by olfactory stimuli, prompting the suggestion of an appropriate range of EEG frequency bands to evaluate psychological states arising from the olfactory stimuli. A novel method, incorporating biological indices and an emotion map, is presented in this research to depict psychological responses to olfactory stimuli in greater detail. Understanding consumer emotional reactions to olfactory products is significantly enhanced by this method, benefiting the areas of product design and marketing.
The convolution module of the Conformer network ensures translationally invariant convolutions, operating uniformly across time and spatial dimensions. The diversity of speech signals in Mandarin recognition tasks is often handled by treating time-frequency maps as images, employing this method. Enfermedad cardiovascular While convolutional networks perform well with local features, dialect recognition demands a comprehensive sequence of contextual information; therefore, this paper presents the SE-Conformer-TCN. Through the strategic insertion of the squeeze-excitation block into the Conformer, the model gains the ability to explicitly represent the relationships between channel features. This subsequently enhances the model's ability to pinpoint pertinent channels, bolstering the weighting of useful speech spectrogram features while diminishing the weighting of less relevant feature maps. Simultaneous implementation of a multi-head self-attention module and a temporal convolutional network is facilitated by incorporating dilated causal convolutions. These convolutions capture spatial relationships within the input time series by scaling the expansion factor and kernel size, ultimately enhancing the model's access to information regarding the positional context within the sequences. Experiments conducted on four public Mandarin datasets show that the proposed model outperforms the Conformer in accent recognition, with a 21% decrease in sentence error rate while maintaining a character error rate of 49%.
Self-driving vehicles are obligated to utilize navigation algorithms for controlling operation in order to ensure the safety of passengers, pedestrians, and other drivers. To attain this target, a critical component is the availability of robust multi-object detection and tracking algorithms. These algorithms provide accurate estimations of the position, orientation, and speed of pedestrians and other vehicles on the roadway. These methods' effectiveness in road driving conditions has not been sufficiently examined in the experimental analyses conducted to date. Within this paper, a benchmark for contemporary multi-object detection and tracking systems is proposed, based on image sequences acquired by a vehicle-mounted camera, utilizing the BDD100K dataset's video data. To assess 22 distinct combinations of multi-object detection and tracking approaches, the presented experimental setup provides metrics that reveal both the successes and limitations of each module within the examined algorithms. Experimental results reveal that combining ConvNext and QDTrack yields the optimal current approach, yet underscore the critical need for significant enhancement in multi-object tracking techniques specifically on images of roads. Our analysis necessitates the expansion of evaluation metrics to incorporate specific autonomous driving features, including multi-class problem formulations and distance from targets, and demands evaluation of method efficacy through simulations of error effects on driving safety.
In many vision-based measurement systems employed in fields like quality control, defect analysis, biomedical imaging, aerial photography, and satellite imagery, the accurate measurement of the geometric characteristics of curved structures in images is of significant importance. This paper endeavors to establish the groundwork for automated vision-based measurement systems dedicated to quantifying curvilinear features, such as cracks present in concrete. Specifically, the aim is to surpass the constraint of employing the widely recognized Steger's ridge detection algorithm in these applications due to the manual determination of the input parameters defining the algorithm, thereby hindering its widespread application in the field of measurement. Transgenerational immune priming This paper presents a method for automating the selection process of these input parameters during the selection phase. A discussion of the metrological effectiveness of the presented approach is provided.