Abstract: Pedestrian detection has so far worked efficiently using four basic important components namely: feature extraction, deformation handling, occlusion handling, and individual or sequential classification proposed in existing methods. This paper has primarily concentrated on collective basic deep learning on each of these factors using and advancing a new deep neural network architecture. In the aforementioned paper, the advanced neural architecture is compared with the current models including the Caltech benchmark dataset and ETH dataset to examine the results and accuracy.