| Crowd-Counting Images in rainy Weather Based on GAN-U-Net model and Multi-column CNN Counting approaches |
| Paper ID : 1121-ICEEM2023 |
| Authors |
|
Rasha ahmed1, Mohamed Abdelazeam2, Prof. Fathi Sayed3, Heba mohamed El-Hoseny *4 1Department of Electronics and Electrical Communications, Obour Higher Institute for Engineering and Technology, Egypt. 2Department of Electronics and Electrical Communications, Faculty of Engineering, Mansoura University, Egypt. 3Department of Electronics and Electrical Communications, Faculty of Electronic Engineering, Menoufia University, Egypt. 4Department of Computer Science, The Higher Future Institute for Specialized Technological Studies, Obour, Egypt |
| Abstract |
| A count of crowds tries to determine the total number of crowds in crowded places in order to avoid disruption of the safety system and maintain crowd safety. Accurately estimating the size of crowds is a challenging process because of numerous challenges like occlusion, complicated backgrounds, varying sizes, uneven distribution, perspective distortion, rotation, different illumination, and changing weather. In this study, we will propose a solution to two crowd count challenges: poor weather (such as rain) and changes in the scale of images. We suggest Crowd-Counting in the rainy Shanghai-Tec image. Based on the GAN-U-Net (de-raining mode) model and the crowd-counting model We eliminate the rain with the U-Net-GAN model and test the results with the two crowd-counting models that have been suggested (SAS-Net [1], SGA-Net [2]). |
| Keywords |
| GAN-U-Net, Crowd Counting, convolution neural network (CNN). |
| Status: Accepted |