Global Journal of Computer Science and Technology, D: Neural & Artificial Intelligence, Volume 22 Issue 1

Global Journal of Computer Science and Technology Volume XXII Issue I Version I 39 ( )D Year 2022 © 2022 Global Journals Neural Network Design using a Virtual Reality Platform convolution neural network to recognize human activities (HAR). In addition, the user can select input data and have them classified by CNN. In addition, we previously created a network to identify human activities (HAR) with the STMicroelectronics AI STM32 module [42]. Inference model a) CNN model classify images from a sample dataset. It is based on the principle of the convolution filter applied to the pixel matrix that composes the image. Convolutional layers produce small transformations on images. Each convolution transforms a section of the image into a single value. The Kernel applied sequentially to the embodiment can highlight the fundamental characteristics of the image and, in this way, produce essential data for the classification of the image itself. Finally, we measure the model's performance through the accuracy and the loss. CNN model def cnnModel(): model = Sequential() # adding the first convolutionial layer with 32 filters and 5 by 5 kernal size, using the rectifier as the activation function model.add(Conv2D(numFilters,(kernalSize1,kernalSize1),input_shape=(numOfRows,numOfColumns,1),activation=' relu')) # adding a maxpooling layer model.add(MaxPooling2D(pool_size=(poolingWindowSz,poolingWindowSz),padding='valid')) # adding a dropout layer for the regularization and avoiding over fitting model.add(Dropout(dropOutRatio)) # flattening the output in order to apply the fully connected layer Convolutional Neural Networks (CNN) represents an artificial neural network architecture that has found wide application in identifying, with a certain probability, the content of the image provided by a computer. From a functional point of view, convolutional neural networks simulate how the human brain processes. Convolutional neural networks are built to analyze images included within certain data sets and classify objects in images within them. Each class of things must have a specific network previously trained for the reconstitution of that class. Finally, they process the data through multiple layers placed between the input and the output. In our study, we have used a CNN to recognize the activities carried out by people daily, such as downstairs, upstairs, sitting, standing, walking. It is mainly used as a supervised learning algorithm and can A Keras model allows building the CNN network, created with the STMicroelectronics AI STM32 and trained on the public dataset Wireless Sensor Data Mining (WISDM). The Keras DL model used is Shahnawax/HAR-CNN-Keras. The code for the model definition is in the following listing: