DRAFT 1
DRAFT 1
Introduction
Hook:
Since many decades ago, the smart city life has always been a dream for majority of the urbanites as it will be a giant leap for human beings if everything in the city could be automatically linked together. With the advancement of computer vision technology, this goal becomes achievable because the computer vision provides accurate analysis and intensive connectivity in many sectors such as transportation, security, medical services and so forth.
General information:
Computer vision is a study of artificial intelligence that focuses on how to enable the machine to analyse and identify object as the same way as human does via digital input such as images and videos. On the other hand, a smart city life is the idea of building an urban area with high operation efficiency by implementing the information and communication technology (ICT) into various aspects of life.
Specific information:
In a smart city, it is essential to utilized Computer vision technology as it acts as the “eyes” of the smart city. With this system adapted inside a smart city, it will help in gathering real time data and to promote public welfare by making the smart city more efficient and a safe place to live.
Thesis statement:
In short, computer vision aid in creating smooth traffic, enhance the security system and help in the medical field in a smart city.
Main Point 1：Overcome Traffic Congestion
First of all, a smart and effective traffic management system can be implemented in a smart city with the help computer vision technology.
S1: Autonomous Vehicles
In fact, an autonomous vehicle is inseparable from the computer vision technology. The computer vision technology gives cars the ability to “see” and allow vehicles to detect the obstacles of surrounding and changings of the environment. Through the deep neural network, the computer vision system will analyze the roads, signs, cars, obstacles, and pedestrians in the on-board camera. For example, the self-driving cars can detect the traffic signs correctly to allow the processor of the cars to automatically run the next instructions. It will definitely improve the ability of autonomous cars and thus reduce the number of accidents. The application of this technology will bring unprecedented travel experience to human beings, reshape the transportation system, and build a true era of intelligent transportation.
S2: Vehicle Identification and Analyze Traffic Condition
Through deep learning, the computer vision can also achieve more dimensional recognition. The current "car face recognition" technology nowadays can not only accurately identify the license plate, but also the color, type, heigh, brand year of the vehicle, the person in the vehicle, vehicle windshield and even special signs or characteristic signs at the rear of the vehicle. This function can be applied to some roads with vehicle height restrictions. After the intelligent vehicle acquire the traffic real-time information, it can change a better travel route early to avoid accidents or traffic congestion. In addition, the vehicle identification can also prevent some crimes from happening. When the computer vision system detects the presence of unknown vehicles, it will automatically increase alertness. The usage of real-time traffic monitoring is not only use to traffic statistics but also be used for forecasting accidents and even be used in the inspection of anti-social behavior in public areas (Barthélemy et al., 2019).
S3: Smart Parking System
Through computer vision technology, it is possible to simulate human visual perception where there is a car parked and which place is empty, directly detect and send the data to the platform, and publish it to the parking lot guidance system. Ho et al. (2019) designed a Computer Vision-Based Roadside Occupation Surveillance System (CVROSS) which can improve the transparency of roadside occupancy and vacancy and reduce the probability of double parking through the parking-space calculation and parking space allocation model.
Main Point 2: Enhancing The Security
Furthermore, computer vision technology is capable in ensuring the efficiency of a smart city security system.
S1: Smart surveillance system
Smart surveillance system has utilized the computer vision technology in automizing the process of crime monitoring. For example, in China, Shen Zhen city have installed more than 200-millions of cameras in detecting the whole city, criminals were detected using computer vision technology. By using this technology jaywalker will immediately decrease their social credit once it is spotted. AI detection is also used to detect the gender, ethnicity, age as well as style of walking of the people. AI detection also used in determine the social credit of a person, which allow the government to punish criminals as well as giving tax break for the person who obey the laws. On the other hand, in US, New Orleans have adopted the idea of using Artificial Intelligence in combating crime, which utilize a real time crime center in monitoring the whole city using cameras. The camera will focus on a certain location when there is some incident going on (By detecting the people that call 911). BriefCam, which able to detect and identify the people that pass through as well as the vehicles. Used computer vision/analytics function in determine unusual behavior of people and vehicles as well as tracking people from other part of the camera by detecting the gentle/clothes of the person. Able to search the people/ object with certain by inputting the characteristic of the object.
S2: Airport security
Not only computer vision technology has utilized in smart surveillance system, it also utilizes in other security system such as in airport security system. In this system, it is capable in detecting the behavior of large group of people such as in detecting the suspect's movement, facial expression and facial recognition. It also enhances the check-point security by auto sensing explosive materials, dangerous weapons and firearms, while ignoring others object that is safe. With this the airport security system that utilized the computer vision technology able to track dangerous weapons and counter terrorism.
S3: Security in business
Lastly, computer vision technology also enhances the security in businesses. For example, in Amazon Go, which is a convenient store that used computer vision technology in detecting the customers, as well as automatically deduct the cost of the products that is take off the shelves by the customers and walk out the store. With this system, it is able to prevent shoplifting because it requires the log in of the users using Amazon Prime in order to enter the store and check out the products. Theft detection system also used by various convenient store by applying advance computer algorithms and technology to help business operator to identify threat and theft based on their unusual behavior during check out period or shopping period.
Main Point 3: Increased Healthcare Quality
It has come to the acknowledgement that computer vision is widely beneficial to the healthcare sector which can contribute to better living standards of smart city li
S1: Aiding doctors in diagnoses
Research done has shown improved accuracy of detection by deep learning algorithms. A test was carried out using a cancer detection model which achieved the area of 94.4% under the Receiver Operating Character (ROC) curve, in contrast to experts which had equivalent or lower performance rate. Increased diagnosis accuracy would prevent unnecessary surgery improving the allocation of hospital resources for more optimal uses.
S2: Application in surgical settings
Computer vision has proved to be a helping hand to doctors in surgical scenes. Computer-vision based algorithms are trained to detect surgeon’s hand movement during operating processes to evaluate surgeons’ skills. Such as in laparoscopic surgery, algorithms are trained to assess surgeon's hands and tool movement according to Global Operative Assessment of Laparoscopic Skills (GOALS) criteria. These algorithms can enhance the surgeon's performance through training and skill assessments, so that surgeons can achieve more precision through computer-aided surgeries, leading to better post-operation results.
S3: Help monitors hospital setting and elderly’s living settings
Research found ways to sustain access to effective elderly healthcare and provide more personal services to the growing demand of elderly healthcare through computer vision-based ambience intelligence. Computer vision monitoring provides non-invasive monitoring of elderly in house settings so that help is never out of reach and prevents any acute health problems, such as a model created to detect falls. Long-term monitoring can provide long-term activity analytics of elderly, such as sleeping, walking and sitting, so that healthcare providers have clearer view of elders’ health in order to formulate better assistance. Monitoring can be done through video recording and thermal video data.
With the growing demands of healthcare services, computer vision helps relieve pressure over healthcare providers so that attention is put where it is most needed. Computer vision can be used for administration tasks such as monitoring protocol or paperwork and release resources for more demanding needs.
Hacking Risks
Kong, Z., Xue, J., Wang, Y., Huang, L., Niu, Z., & Li, F. (2021). A survey on adversarial attack in the age of artificial intelligence. Wireless Communications and Mobile Computing, 2021. https://doi.org/10.1155/2021/4907754
Kwon, H., & Lee, J. (2021). Diversity adversarial training against adversarial attack on deep neural networks. Symmetry, 13(3). https://doi.org/10.3390/sym13030428
Park, S., & So, J. (2020). On the effectiveness of adversarial training in defending against adversarial example attacks for image classification. Applied Sciences (Switzerland), 10(22), 1–16. https://doi.org/10.3390/app10228079
Sharif, M., Bhagavatula, S., Bauer, L., & Reiter, M. K. (2016). Accessorize to a crime: Real and stealthy attacks on state-of-the-art face recognition. Proceedings of the ACM Conference on Computer and Communications Security, 2016, 1528–1540. https://doi.org/10.1145/2976749.2978392
Xu, H., Ma, Y., Liu, H. C., Deb, D., Liu, H., Tang, J. L., & Jain, A. K. (2020). Adversarial attacks and defenses in images, graphs and text: A review. International Journal of Automation and Computing, 17(2), 151–178. https://doi.org/10.1007/s11633-019-1211-x