In systems, accurate detection of people can be difficult, especially in dense crowds where individuals could be blocked partially or completely for any time duration. Human detection is essential in various applications such as abnormal event detection, human stride classification, crowd analysis, human recognition, gender categorization, and fall detection for the elderly. In this research paper, with the help of the Tiny-YOLOv3 and SSD algorithms, we detect humans in crowded images under different conditions. These algorithms divide the photo into areas and predict bounding bins and chances for every region. The projected probabilities weigh these bounding boxes and make their detection based on the final weights. Models are trained using COCO dataset. Tiny-Yolov3 and SSD models evaluate the performance for different images under various conditions, and it has been found that the Tiny-Yolov3 algorithm has performed better with the highest accuracy of 0.80 as compared to the SSD algorithm, which has the highest accuracy of 0.50. The Tiny-YOLOv3 algorithm has a precision of 1, which is the highest among all fed im-ages for Tiny-Yolov3, whereas SSD has a highest precision of 0.75.