|
Document Type |
: |
Thesis |
|
Document Title |
: |
CNN-Based Crowd Counting Through IoT: Application For Saudi Public Places عد الناس في البيئات المزدحمة باستخدام الشبكة العصبية التلافيفية وانترنت الأشياء: تطبيق للأماكن العامة في المملكة العربية السعودية |
|
Subject |
: |
Faculty of Computing and Information Technology |
|
Document Language |
: |
Arabic |
|
Abstract |
: |
Crowd counting in specific places has lately been considered as a significant contribution in many applications in terms of security and economic values. Recently, the Kingdom of Saudi Arabia has considered new ways and methods to diversify sources of income, where many events hosted newly though the entertainments sector such as sports, concerts, festivals, and exhibitions. To ensure security, comfort, and safety of the visitors, crowds should be managed and estimated. Crowd counting is a very beneficial application not only to help in resolving security and safety problems, but also it plays a significant role in reducing waiting time for visitors, by giving indicators, projections and advice on crowded places. In this thesis, a mobile-based model is proposed for counting people in high and low crowded public places in Saudi Arabia under various scene conditions with no prior knowledge. The proposed Deep CNN model (DCNN) is built based on the structure of the convolutional neural network (CNN) with small kernel size and two fronts. A convolutional neural network (CNN) as the front-end and a multi-column layer with Dilated Convolution for the back-end of CNN to increase the efficiency of the training model. Our model is an easy-to-trained model because of its simple convolutional structure. In addition to the improvement of efficiency, the proposed method accepts images of arbitrary sizes/scales as inputs from different cameras. The applicability of the proposed method has been evaluated by incorporating IoT architecture, where virtual cameras connected to the Internet to capture live pictures of different public places and it could be generalized to surveillance camera as well. To evaluate the proposed approach, a new dataset is developed. It contains Saudi people images of traditional and non-traditional Saudi uniforms. The approach is also trained and tested on some challenging existing dataset. The result shows that our method significantly improves efficiency by reducing the error rate over the existing counting methods. |
|
Supervisor |
: |
Dr. Salma Mohamad Kammoun |
|
Thesis Type |
: |
Master Thesis |
|
Publishing Year |
: |
1440 AH
2019 AD |
|
Co-Supervisor |
: |
Dr.Manar Sayed Salama |
|
Added Date |
: |
Wednesday, August 21, 2019 |
|
Researchers
| مها حمدان العتيبي | Al-Otibi, Maha Hamdan | Researcher | Master | |
|