M2_1_Written Research Assignment 1

docx

School

George Mason University *

*We aren’t endorsed by this school

Course

524

Subject

Electrical Engineering

Date

Apr 3, 2024

Type

docx

Pages

Uploaded by JusticeSkunkPerson786

Written Research Assignment Question No: 01 I. The research question formulated as a hypothesis or question. AI techniques for identifying Fake and Real Faces II. One paragraph description of the research question providing the context and more detailed explanation. Intelligent solutions are proposed by artificial intelligence to address issues in the real world. Deep learning, neural networks, and other artificial intelligence techniques have the capacity to carry out tasks that are capable of being carried out by human intelligence. Even if AI techniques are doing marvels in executing many tasks, human performance is still unmatched. Finding false online content is an excellent example of some of the tasks which AI has done. In today's world, being able to spot false films of anyone is essential. Some of Deep Fakes' fake videos are quite well-done. These methods enable the creation of genuine facial expressions from a typical facial photograph. Due to this, fraudulent news and messages are being spread online, and many people are falling for it. There are techniques to counter this chaos AI is doing tremendous job in identifying the pictures with fake faces. It will help labelling the false information on the internet. III. One paragraph in which you explain why you consider that your question is not too narrow or too general. The question is surely not narrow as it will discuss various techniques for identifying fake faces and will ultimately discuss the best approach. Also, it is not too general as I will not be working on all the fake content which is there on the internet. The question is focused on identifying the fake faces only IV. One paragraph in which you explain why you consider that your question is not too simple (obvious, not much to do to answer it) or too complex (requiring a lot of time to analyze) This question is not too simple as it cannot be answered by a single line. One needs to do research on various techniques and discuss them in the answer. Also, it pushes you to identify the best technique as well. It also does not restrict you from developing the best technique so one can go towards building the best algorithm for identifying fake faces.

V. A list with at least 3 relevant and authoritative references related to the proposed research question (1 point) 1) Yang, Jiachen, et al. “Detecting Fake Images by Identifying Potential Texture Difference.” Future Generation Computer Systems, vol. 125, Dec. 2021, pp. 127–135, 10.1016/j.future.2021.06.043. Accessed 22 Nov. 2021. 2) Frank, Joel, et al. “Leveraging Frequency Analysis for Deep Fake Image Recognition.” Proceedings.mlr.press, PMLR, 21 Nov. 2020, proceedings.mlr.press/v119/frank20a. Accessed 14 Sept. 2022. 3) Mo, Huaxiao, et al. “Fake Faces Identification via Convolutional Neural Network.” Proceedings of the 6th ACM Workshop on Information Hiding and Multimedia Security, 14 June 2018, 10.1145/3206004.3206009. Question No: 02 I. The research question formulated as a hypothesis or question. Identifying Anomaly in Surveillance Videos for real-time monitoring II. One paragraph description of the research question providing the context and more detailed explanation. Data about a wide range of anomalies can be easily obtained from surveillance cameras. This information will include a range of videos for unusual situations. Numerous neural network topologies for developing a reliable model that can quickly detect abnormalities have been made available via deep learning. Law enforcement organizations can help to reduce the damage by using real-time anomaly detection. Anomaly detection has been transformed by artificial intelligence, much like every other field. As a result, this subject is the subject of extensive investigation. Finding the best ways to combine several methodologies can produce an effective and reliable anomaly detection system. III. One paragraph in which you explain why you consider that your question is not too narrow or too general. It is not a narrow question as it does not limit the research to one technique. It allows research about various techniques and identifying the best solution. Also, it will cover a variety to anomalies like Abuse, Assault, Explosion or fighting. It is also not too general as it specifically asked about anomaly detection in surveillance videos and not any type of videos.

IV. One paragraph in which you explain why you consider that your question is not too simple (obvious, not much to do to answer it) or too complex (requiring a lot of time to analyze) It is not a simple question because it is asking about ways of identifying anomalies. There are various ways of identifying anomalies and doing research on the available ways. This will require a lot of research therefore it is not a simple question to answer. V. A list with at least 3 relevant and authoritative references related to the proposed research question 1) Sultani, Waqas, et al. “Real-World Anomaly Detection in Surveillance Videos.” Openaccess.thecvf.com, 2018, openaccess.thecvf.com/content_cvpr_2018/html/Sultani_Real- World_Anomaly_Detection_CVPR_2018_paper.html. 2) Doshi, Keval, and Yasin Yilmaz. “Online Anomaly Detection in Surveillance Videos with Asymptotic Bound on False Alarm Rate.” Pattern Recognition, vol. 114, June 2021, p. 107865, 10.1016/j.patcog.2021.107865. Accessed 21 Mar. 2021. 3) Murugesan, M., and S. Thilagamani. “Efficient Anomaly Detection in Surveillance Videos Based on Multi Layer Perception Recurrent Neural Network.” Microprocessors and Microsystems, vol. 79, Nov. 2020, p. 103303, 10.1016/j.micpro.2020.103303. Accessed 2 Dec. 2020.

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version