RIS Based Hand Gesture Recognition DatasetOverviewThis dataset contains images for gesture recognition, divided into two main sets: dataset0608 and data_synthetic_variab. The data was collected using a wooden hand.    dataset0608This dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized). This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame_{i}{posture}{n_med}' - ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}{n_med}'  For each gestures = {close, two, open}, we have n_med values from 0 to 114 and 10 frames. Therefore, the ris_random and ris_optimized folders contain 10 frames × 115 measurements × 3 gestures = 3450 files, while the ris_random2 and ris_optimized2 folders contain 1 × 115 measurements × 3 gestures = 345 files. data_synthetic_variabThis dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized).  This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame{i}{posture}{n_med}' - ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}_{n_med}'  For each gestures = {close, two, open},  we have n_med values from 0 to 8 and 10 frames. This dataset provides additional synthetic data with variations in hand position to increase the dataset's diversity. Each gesture is represented by 8 different ways, where the hand position was slightly modified between each sample. These real data were used as a basis for generating synthetic data. By using the functions in the files "multiply_files.txt" and "add_gaussian_noise.txt," the dataset was expanded and made more realistic by adding Gaussian noise to the images. Therefore, the ris_random and ris_optimized folders contain 10 frames × 8 measurements × 3 gestures = 240 files, while the ris_random2 and ris_optimized2 folders contain 1 × 8 measurements × 3 gestures = 24 files. Functions* add_gaussian_noise.txt: This script adds Gaussian noise to the images to simulate real-world conditions and improve the robustness of the model.* compact_files_frames.txt: This script combines multiple frames into a single image, which can be useful for certain types of analysis.

RIS Based Hand Gesture Recognition DatasetOverviewThis dataset contains images for gesture recognition, divided into two main sets: dataset0608 and data_synthetic_variab. The data was collected using a wooden hand.    dataset0608This dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized). This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame_{i}{posture}{n_med}' - ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}{n_med}'  For each gestures = {close, two, open}, we have n_med values from 0 to 114 and 10 frames. Therefore, the ris_random and ris_optimized folders contain 10 frames × 115 measurements × 3 gestures = 3450 files, while the ris_random2 and ris_optimized2 folders contain 1 × 115 measurements × 3 gestures = 345 files. data_synthetic_variabThis dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized).  This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame{i}{posture}{n_med}' - ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}_{n_med}'  For each gestures = {close, two, open},  we have n_med values from 0 to 8 and 10 frames. This dataset provides additional synthetic data with variations in hand position to increase the dataset's diversity. Each gesture is represented by 8 different ways, where the hand position was slightly modified between each sample. These real data were used as a basis for generating synthetic data. By using the functions in the files "multiply_files.txt" and "add_gaussian_noise.txt," the dataset was expanded and made more realistic by adding Gaussian noise to the images. Therefore, the ris_random and ris_optimized folders contain 10 frames × 8 measurements × 3 gestures = 240 files, while the ris_random2 and ris_optimized2 folders contain 1 × 8 measurements × 3 gestures = 24 files. Functions* add_gaussian_noise.txt: This script adds Gaussian noise to the images to simulate real-world conditions and improve the robustness of the model.* compact_files_frames.txt: This script combines multiple frames into a single image, which can be useful for certain types of analysis.

In recent years, the utilization of Radio Frequency (RF) signals for Human Posture Recognition (HPR) has emerged as a promising approach in wireless sensing technology. Recent studies within RF systems have demonstrated the effectiveness of S21 parameters for human body-related classification tasks. Inspired by these advancements and the enhanced network performance offered by WiFi-6E, our system leverages S21 parameters within this frequency range for posture recognition. By analyzing the S21 parameters, we introduce SpecRF-Posture, a novel system designed for the accurate classification of human posture. Unlike previous approaches, we explore the use of low-cost hardware, based only on passive specular reflections that occur in the path between a steerable horn transmitter, a reflective surface, the Space-of-Interest (SoI), and an omnidirectional receiver. To characterize the region where the individual is positioned using specular reflections, our system achieves beam scanning by mechanically rotating the transmitter at regular intervals. Our work evaluates the viability of utilizing passive scatters within the propagation medium for HPR or similar tasks, at a low hardware cost.For each posture (Standing, T-shape, Side, Sitting, Lying down), we collected 120 samples (referred to as sweeps) from one volunteer, resulting in a dataset of 600 samples (1 subject × 5 postures × 120 sweeps). Each sample is characterized by dimensions (n_resolution, n_angles), where n_resolution=151 represents the number of frequencies between 5.925 GHz and 6.5 GHz, defining the resolution of the frequency range, and n_angles = 25 denotes the number of angles through which the antenna rotates. This spectrum of postures included minor movements of the hands or head to introduce variability into the data and make the model robust to different ways of assuming these postures.Each sweep corresponds to a .csv file where each row corresponds to a frequency and each pair of columns corresponds to the real and imaginary parts of the S21 parameter measured at each angle of rotation of the transmitter antenna.

In recent years, the utilization of Radio Frequency (RF) signals for Human Posture Recognition (HPR) has emerged as a promising approach in wireless sensing technology. Recent studies within RF systems have demonstrated the effectiveness of S21 parameters for human body-related classification tasks. Inspired by these advancements and the enhanced network performance offered by WiFi-6E, our system leverages S21 parameters within this frequency range for posture recognition. By analyzing the S21 parameters, we introduce SpecRF-Posture, a novel system designed for the accurate classification of human posture. Unlike previous approaches, we explore the use of low-cost hardware, based only on passive specular reflections that occur in the path between a steerable horn transmitter, a reflective surface, the Space-of-Interest (SoI), and an omnidirectional receiver. To characterize the region where the individual is positioned using specular reflections, our system achieves beam scanning by mechanically rotating the transmitter at regular intervals. Our work evaluates the viability of utilizing passive scatters within the propagation medium for HPR or similar tasks, at a low hardware cost.For each posture (Standing, T-shape, Side, Sitting, Lying down), we collected 120 samples (referred to as sweeps) from one volunteer, resulting in a dataset of 600 samples (1 subject × 5 postures × 120 sweeps). Each sample is characterized by dimensions (n_resolution, n_angles), where n_resolution=151 represents the number of frequencies between 5.925 GHz and 6.5 GHz, defining the resolution of the frequency range, and n_angles = 25 denotes the number of angles through which the antenna rotates. This spectrum of postures included minor movements of the hands or head to introduce variability into the data and make the model robust to different ways of assuming these postures.Each sweep corresponds to a .csv file where each row corresponds to a frequency and each pair of columns corresponds to the real and imaginary parts of the S21 parameter measured at each angle of rotation of the transmitter antenna.

The dataset will be uploaded to this path after the study is accepted for publication. The dataset consists of training and test data and label matrices for single-pixel compressive DoA estimation for mmWave metasurface.