| Autor: |
Borkar, Nishant M., Dhoot, Manali, Bisani, Piyush, Agrawal, Sanskar, Yadav, Shubham |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2024, Vol. 3131 Issue 1, p1-8, 8p |
| Abstrakt: |
Recently, some academics have become interested in subjects like crowd estimation from still images, public safety, and crowd density counts. Crowd counting algorithms use a large number of programmed regressions to get the crowd size. There are several ways to gauge the size of a crowd; however most of them are constrained by specific limitations: They can only handle groups of up to ten individuals and cannot control crowds of hundreds or thousands. They should not be used to count individuals, just to measure crowd density. Only crowd films will be effective in place of still images. In this article, we investigate how to precisely count people and estimate their density in still photographs using deep learning. In this article, we investigate how to precisely count people and estimate their density in still photographs using deep learning. The count is calculated using the CNN (Convolutional Neutral Networks) methodology employing a dataset and data processing. There are many layers in CNN technique. We have applied this technique in both sparse and dense image and we can get the approximate count. It is very useful in security as there are many concerts or public events so we can get the head count of humans. Efficiency of proposed approaches is demonstrated in many experiments. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
