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 35 ( )D Year 2022 © 2022 Global Journals Neural Network Design using a Virtual Reality Platform solution for image recognition [27]. They are built to analyze images included within certain data sets and classify objects in images within them. CNN is a network composed of several convolutional layers [28]. Each processing layer comprises a convolutional filter, an activation function (ReLu), a pooling function, and a fully connected layer. At the end of each processing step, an input is generated for the next layer. In the convolutional operations, the trained filter set is convolved with input images to extract the specific feature to create the feature map, which becomes the input for the next filter. Finally, the design of a CNN network requires a training period followed by a test phase. During the training phase, the images are labeled and transferred to the subsequent layers to allow the structure to convert from the representation level of the original input to a higher level and more abstract representation to constitute the reference feature maps with which the network must compare the output feature maps. Once the training and test phase of the network has been completed, we will determine the survey's accuracy level. Each layer comprises three levels: Convolution, ReLu, and Pooling. − Convolutional Level (CONV) is the main level of the network. Its objective is to identify patterns. They are multiple, designed to identify features present in the initial image. Each layer learns to extract specific parts of the photos placed at its entrance. Multiple layers in cascade combine the features of the previous layers with higher programmed extraction levels. − Rectified Linear Unit (ReLu) Level is placed after the convolutional level and can cancel negative values obtained in the previous classes. − Pool level allows identifying if the study characteristic is present in the previous story. The pooling layer obtains images with a particular resolution at the input and returns the same number of pictures with fewer pixels. The result of the convolution operations is the production of feature maps obtained with the help of filters that are matrices containing proper values for finding specific characteristics in the input images. At the end of the sequence of convolutional layers, there is then the fully connected level (FC) which aims to establish the identifying classes obtained in the previous levels according to a certain probability. Each category represents a possible answer that the system will most likely choose. During the recognition phase, the network performs a classification operation to identify which class the input image belongs to, identifying the one with the highest probability. The values of the filters are initially chosen randomly. They are subsequently improved at each iteration of the training phase. We estimate that the model's predictions are plausible; in practice, we measure the discrepancy between actual and predicted values. The error is subsequently processed using the stochastic gradient technique. It is a cyclical technique consisting of two phases: * backpropagation; * updating the gradient value. After propagating the forward phase during training, the outputs are produced to determine the prediction error compared with the expected ones. This error is used to calculate the gradient of the loss function. The backward propagation phase then sends the error through the network layers and updates the weights using the stochastic gradient descent to improve the network's performance on the activity it is trying to learn [29]. IV. V irtual R eality The virtual environment must have the following characteristics: • Perception of really being in that world. The use of special equipment amplifies this feeling: software capable of reproducing 3D environments, a virtual reality viewer, integrated audio systems that offer surround support. • The possibility of interacting with movements of the body, head, and limbs increases the feeling of taking possession of that dimension. For example, cyber-gloves, virtual limbs, joypads, etc., allow the user to touch, move, manipulate, or make virtual objects as if they were real. Virtual reality technology refers to a computer- generated virtual world [30]. To immerse yourself in this multidimensional world, the user must wear a helmet containing a display that incorporates a sensor to track the movement and position of the wearer (HMD - Head Mounted Display). The sensors detect the position and movement data, which are used to update the image of the virtual world. In this way, the user is projected into this virtual world and can interact with the objects present in it, for example, simulating carrying out usual activities or driving a car or running. By manipulating three variables (space, time, and interaction) and the availability of a graphic interface, it is possible to create a dimension characterized by a strong sense of reality, whereby the subject believes he is in that world and can interact with it. The relationship between presence and immersion gives cognition in virtual reality. The first term refers to the level of psychological realism that a subject experiences from interacting with the virtual world. From the wise point of view, the second term means the ability of the environment to involve the senses of the subject, isolating it from the stimuli of the real environment [31].