Each user has 150 samples analyzed by the hand gesture recognition model. The sample responses for the fields mentioned above ( vectorOfLabels, vectorOfTimePoints, vectorOfProcessingTimes, class) are stored from idx_0 to idx_149. VectorOfProcessingTime: Vector with the processing times corresponding to the analysis of each sliding window.Ĭlass: The most frequent label (mode) in the vectorOfLabels.
VectorOfTimePoints: Vector with the time corresponding to the predicted labels of vectorOfLabels. VectorOfLabels: Vector with the predicted labels for the sliding windows of each sample. The responses corresponding to each user are divided as shown below: The “responses.json” file has the following structure: json file with the responses of each of the users.
You can observe the progress of the code by observing the command window.Īt the end of analyzing all users, the code automatically generates a. The code starts running and analyzes all users.
In our case, we classified six gestures corresponding to 306 testing users. In the variable userFolder you can change for testing or training to choose the user group. Run the script main.m After running this script, you will have to wait several minutes to obtain the results. You need to do this step only once in the computer where you will run the code.ĭownload the dataset before running the code from the following link:Ĭopy and replace the folders corresponding to the dataset in the downloaded repository folder. Open Matlab and choose the example folder.Ĭompile the mex function that computes the DTW distance by running the script compileDTWC.m which is in the folder /DTW distance. Go to our GitHub repository and download or clone the example to manage the dataset in Matlab Example. If you want to use the dataset and the code for commercial purposes, please contact to the correspondent author of the paper, Marco E. Each script contains a description of its function as well as the copyright information. The “Hand Gesture Model Example MATLAB” folder contains the Matlab code to manage the EMG Database. We implemented a Real-Time Hand Gesture Recognition based on Artificial Feed-Forward Neural Networks to test the data of each user. This document describes briefly the steps needed to run the Matlab code to manage our proposed EMG Database.
0 kommentar(er)
