Hidden Layer Reshaping on Convolutional Neural Network

Artificial Intelligence (AI) technologies have become really popular in recent years. From ChatGPT to Tesla cars, many applications can benefit from these type of technologies. Automotive, healthcare, biomedical, cybersecurity, finances, and retail are some of the fields that take advantage of it. It has been seen that AI can solve complex problems, but there is still work to be done to optimize it. A deep learning neural network (DLNN) tries to simulate how a human brain operates. These DLNNs are made up of artificial neurons which are connected by weight that are modified when the network is trained. These networks are used to create AI technologies like the ones mentioned before. Computer vision is another of those areas that benefits greatly and it is the focus of this research. Computer vision tries to mimic human vision by enabling computers, devices or machines to understand, interpret or manipulate what is being seen. DLNNs help with tasks such as classification, object detection, tracking, and image manipulation which are all part of computer vision. The research focuses on optimizing a Convolutional Neural Network (CNN). These types of networks are commonly used when the input data for the neural network are images. The network filters the images and down samples them to make it more manageable when handling the large amount of data incoming from them. The specific layer of interest to improve is the hidden layer. There is not a set number of neurons to be used for this layer, so different approaches will be used to compare accuracy and training time of the network.