<p>Detecting pedestrians is an essential problem in computer vision with a wide range of applications, including self-driving cars, surveillance, and human-computer interaction. Here, we proposed a novel Mask Region-based Convolutional Neural Network (Mask R-CNN). Approach to achieve Pedestrian detection using the Inception ResNetV2 architecture. The innovation is the plug-and-play of the Inception ResNetV2 backbone, which has been very effective at extracting finer image features into Mask R-CNN, a framework for object instance segmentation. The effectiveness could lead to a composite detection and segmentation model of a pedestrian in complex situations where people are masked or jammed together with other objects. Using the Inception ResNetV2 architecture, which provides a backbone of hierarchical, multi-scale features to represent fine details and contextual information in input images, has many advantages for improving the model’s representation capacity. This leads to better performance in pedestrian detection, especially for overlapping pedestrians. In addition, the Mask R-CNN framework enhances this by providing accurate instance segmentation, specifically indicating pedestrian contours. Using a well-established benchmark dataset for pedestrian detection, the proposed approach outperforms state-of-the-art results in object detection and instance segmentation across comprehensive experiments.</p>