Multi Object Detection Based on Deep Learning and Discrete Wavelet Transform

Object detection is one of the most important applications in the field of artificial intelligence, as is the case in identifying faces and li-cense plates. As for the developed technology, work is applied with the help of wavelets to improve the results. In this work, new Discrete Her -mite wavelets Transform (DHWT) were derived from Hermite poly-nomials (HP) so that their special role in this work is to process the input image in order to analyze the image, remove noise and com-press the image in order to reach reading the most important image quality criteria, peak signal-to-noise ratio (PSNR) and average error square ratio (MSE), compression ratio (CR) and bits per pixel (BPP) to be pre-processed stage. After that, the deep learning phase begins with the creation of the convolutional neural network with the help of the new wavelets, Discrete Wavelet Convolutional Neural Network (DWCNN) in order to set the most important points of the object to be detected in order to reach the most important results.