Signal Processing
JVNV is a Japanese emotional speech corpus with verbal content and nonverbal vocalizations whose scripts are generated by a large-scale language model.
Existing emotional speech corpora lack not only proper emotional scripts but also nonverbal vocalizations (NVs) that are essential expressions in spoken language to express emotions.
We propose an automatic script generation method to produce emotional scripts by providing seed words with sentiment polarity and phrases of nonverbal vocalizations to ChatGPT using prompt engineering.
- Categories:
671 Views
The simulation experiment is based on Candence 16.6 software, where the tolerance of the resistance (R) is set to 5%, the tolerance of the capacitance (C) is set to 10%, the input is a single-pulse signal (amplitude 5 V, pulse width 10 µs, eriod 2ms), and the working temperature is set to 27 ℃. The operational amplifier(op-amp) uses the actual UA741 pspice model. The experiment includes the soft fault diagnosis of Sallen-Key band-pass filter circuit (TC1), Four-op-amp biquad high-pass filter circuit (TC2), and Leap-frog low-pass filter circuit (TC3).
- Categories:
731 Views
In this brief, the distributed cubature information filtering method is proposed to solve the state estimation problem of target in passive sensor network. Firstly, the observation system model of bearing-only sensor network is established and analysised. The sensor node pairs only measure the relative angle information, and then the state estimation of the target is realized based on the DCIF algorithm.
- Categories:
63 Views
Our dataset consists of a pretraining dataset and a fine-tuning dataset. The pretraining dataset is generated using the FDTD method. We simulated a scenario for underground pipeline detection, where the transmitter (tx) is located above the ground, and the receiver (rx) is approximately 30 cm away from the transmitter. The target pipeline buried underground has depths ranging from 1 to 3 meters and a length of approximately 10 meters.
- Categories:
203 Views
Our dataset consists of a pretraining dataset and a fine-tuning dataset. The pretraining dataset is generated using the FDTD method. We simulated a scenario for underground pipeline detection, where the transmitter (tx) is located above the ground, and the receiver (rx) is approximately 30 cm away from the transmitter. The target pipeline buried underground has depths ranging from 1 to 3 meters and a length of approximately 10 meters.
- Categories:
89 Views
This dataset is recorded by 26 subjects in Shandong Provincial Hospital using wearable ECG devices. It totally includes 208 segments with a duration of 30 seconds. The sampling rate is 256Hz. All the data format is ‘.mat’. This dataset can be used for signal quality assessment as the unacceptable category. All the data are recorded in free-living conditions with various noises. This dataset is recorded by 26 subjects in Shandong Provincial Hospital using wearable ECG devices. It totally includes 208 segments with a duration of 30 seconds. The sampling rate is 256Hz.
- Categories:
375 Views
The Numerical Latin Letters (DNLL) dataset consists of Latin numeric letters organized into 26 distinct letter classes, corresponding to the Latin alphabet. Each class within this dataset encompasses multiple letter forms, resulting in a diverse and extensive collection. These letters vary in color, size, writing style, thickness, background, orientation, luminosity, and other attributes, making the dataset highly comprehensive and rich.
- Categories:
517 Views
This dataset presents a collection of real-world RF signals encompassing three prominent wireless communication technologies: Wi-Fi (IEEE 802.11ax), LTE, and 5G. The data aims to facilitate advanced research in spectrum analysis, interference identification, and wireless communication optimization. The signals were meticulously captured under varying conditions to ensure a broad representation of real-world scenarios, including different modulation schemes, channel conditions, and data rates.
- Categories:
3318 Views
Wearable and low power devices are vulnerable to side-channel attacks, which can retrieve private data (like sensitive data or the private key of a cryptographic algorithm) based on externally measured magnitudes, like power consumption. These attacks have a high dependence on the data being encrypted -- the more variable it is, the more information an attacker will have for performing it. This database contains ECG data measured with a wearable sensorized garment during different levels of activity.
- Categories:
376 Views
To address the current problem of decoupling between joints and the fixed synergy relationship of underactuated fingers, an adaptive and dexterous underactuated finger named the LMH finger was designed in this paper.
- Categories:
150 Views