Crowd Counting Using Region Convolutional Neural Networks

Monitoring the number of people is essential to estimate the level of crowds in a public area, especially during this Covid19 pandemic. CCTV recording needs to process for counting the number of people in a crowd at a specific time. However, counting people on CCTV is not easy. It can be approached by detecting a specific object from a compilation of frames with a certain size that makes up the image. This study proposed the Faster Region-Convolutional Neural Networks (Faster R-CNN) method with ResNet50 to count the number of people in a crowd from the low-resolution image from CCTV. The research gave that crowd counting with the Faster RCNN needs consideration to choose appropriate architecture. ResNet50 architecture provided an accuracy of 97.20% in detecting the number of people in the crowd image. It was compared to other detectors based on previous studies with the same dataset and gave the highest accuracy. Region Proposal Networks makes Faster RCNN robust. Although the various number of people in the crowd image, quality of the dataset, and anchor aspect ratio values provide good results improve accuracy. Besides, the appropriate learning parameters make the method performance more optimal. This configuration can be applied to real-time testing so that it gave the best results of 86% using Faster RCNN and ResNet50.

Face Mask Detection using Convolutional Neural Network

During pandemic CoVID 19, people must use face masks in public areas to prevent and reducing the risk of transmission and spread of the virus. Computer Vision can help to monitor the use of face masks based on images captured via CCTV. Several public areas have installed CCTV that can monitor using masks, but too many people in the area would create problems. Face and side masked face detection is a challenge, given the removal of facial features such as the mouth and nose. A previous study built a mask detection system using Convolutional Neural Networks (CNN) based models, which produced high accuracy but was limited to the front face. This research proposed the CNN method to detect masks based on facial images taken from cameras in public areas. Images containing faces from CCTV are segmented, each faces first using the Retina Face. Experiments were carried out on a single face image in mask detection, resulting in an accuracy of 97.33%. These excellent results are not surprising given CNN's ability to recognize patterns. The most important thing is the segmentation of the face region from one image, which is then tested to produce an accuracy of 82.46%. We selected the best configuration from the two experiments, combined into a mask detection from an image containing multiple faces. The results also showed a significant effect between the face detection method and the learning rate value on the accuracy of the mask use detection system, with the best results of 79.45% using the RetinaFace face detection model. © 2021 IEEE.