ICACT20220317 Question.1
Questioner: pdkg@gmail.com 2022-02-13 오후 3:15:22 |
ICACT20220317 Answer.1
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-13 오후 3:15:22
Chrome  Click!! |
| What are the main attributes of this approach? |
Regularly, to retrieve and explore content in videos, the traditional solution is that we oblige to use the human eyes and examine those videos manually until we obtain the content that appears in the video. Our architecture provides a solution to retrieve content in video, finding content in video in a simple, efficient way. Using this solution, camera system administrators can easily finding video content, thereby saving costs and human resources to operate related systems. |
ICACT20220317 Question.10
Questioner: ruigelwang@163.com 2022-02-15 오후 9:33:50 |
ICACT20220317 Answer.10
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오후 9:33:50
Chrome Click!! |
| Have you applied the conceptual architecture for deep-learning-based video-objects retrieving system in actual scenarios? What was the effect of video-objects retrieving system? |
|
ICACT20220317 Question.11
Questioner: namacabale@gmail.com 2022-02-16 오후 12:33:41 |
ICACT20220317 Answer.11
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-16 오후 12:33:41
Chrome Click!! |
| Dear Author, you have a very good research. Is there an existing technology almost similar to your proposal that has been deployed in real world environment? What does it lack that your work specifically addressed? Thank you and stay safe. |
|
ICACT20220317 Question.12
Questioner: quandh@vnu.edu.vn 2022-02-16 오후 10:39:12 |
ICACT20220317 Answer.12
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-16 오후 10:39:12
Chrome Click!! |
| Thanks for your work and your presentation. I am wondering if the method can be available in real-time scenario. Can you please provide the current fps of the video once the detection method is applied? |
|
ICACT20220317 Question.2
Questioner: tomayoon@ieee.org 2022-02-14 오후 2:06:09 |
ICACT20220317 Answer.2
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-14 오후 2:06:09
Chrome Click!! |
| First of all, thank you for the good research paper and presentation contents. I would like to know more about the research environment used in this study, and also in detail about the equipment & operating systems, development languages and software libraries you used. |
In this study, we used some frameworks and open-source tools to create the video-objects retrieving system. i.e., Python IDE, Django framework for creating web runtime environment, Yolo algorithm version 5 for detecting objects in videos. The system can run on both MS Windows and Linux OS. |
ICACT20220317 Question.3
Questioner: faryal.nosheen@ceme.nust.edu.pk 2022-02-14 오후 2:08:44 |
ICACT20220317 Answer.3
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-14 오후 2:08:44
Chrome Click!! |
| I like to appreciate your work, it really matches present era needs. There are many algorithms available for object detection, why you choose Yolo algorithm |
Yes, as you said, there are many algorithms available for object detection, currently. Yolo was chosen because it requires lower computing hardware while yielding good accuracy results. |
ICACT20220317 Question.4
Questioner: mots@etri.re.kr 2022-02-14 오후 2:12:51 |
ICACT20220317 Answer.4
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-14 오후 2:12:51
Chrome Click!! |
| Thanks for the presentation. What are the minimum system requirements to run the system you designed? |
I didn't check the minimum system requirements but we can install and run the system with a normal Desktop PC (CPU core i5, 4Gb of RAM, without Graphic card).
The more powerful computer, the more system detects objects faster since it uses the GPA for matrix multiplication.
|
ICACT20220317 Question.5
Questioner: jongjcho@korea.ac.kr 2022-02-15 오전 9:39:41 |
ICACT20220317 Answer.5
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오전 9:39:41
Chrome Click!! |
| Thanks for your awesome presentation. since the development of CCTV, it can cause a significant privacy breach, I am wondering how can you protect these data from attack. thanks for your answer! |
Thanks for your interesting question. To protect cctv video data, we have some different solutions: - User can only connect to the system through LAN or VPN. - Blur video content when viewing outside of the allow area. - Access only for connection with pre-registered devices. |
ICACT20220317 Question.6
Questioner: shenhanji@cnic.cn 2022-02-15 오전 11:09:19 |
ICACT20220317 Answer.6
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오전 11:09:19
Chrome Click!! |
| What is your experimental data set? What problems does your research solve in the real environment? |
We introduced the architecture to create a video retrieval system. No matter what the data set is, our system can use for finding and retrieving content from CCTV videos data. Please watch the demonstration of the system in the clip uploaded on Youtube (at 10:30 in the clip) |
ICACT20220317 Question.7
Questioner: rna0415@g.skku.edu 2022-02-15 오후 3:05:23 |
ICACT20220317 Answer.7
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오후 3:05:23
Chrome Click!! |
| Thank you for your work! It was an interesting and useful idea in the physical world! My question is that you said the solution you proposed promises to bring high efficiency in finding video content, but how do you demonstrate that result is efficient? I would like to see the comparison with baseline (metric or performance) for the concrete proposal. Thank you in advance! |
Thank you for your question! In the current system, we use the YOLO v5 for detecting objects, you can see that the average accuracy of predicting results of YOLO v5 is about 60% to 80%. The efficiency of our system goes with the accuracy of the current YOLO version. Our system uses deep-learning-based algorithms and YOLO is a case study we applied. Our idea is one of the new approaches for retrieving content from video files, in my opinion. For the concrete proposal, you can see the results in the demostration video. |
ICACT20220317 Question.8
Questioner: cho_michael_yoshi@sec.ee.kagu.tus.ac.jp 2022-02-15 오후 10:01:44 |
ICACT20220317 Answer.8
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오후 10:01:44
Chrome Click!! |
| Thanks for the presentation.
For applications such as security, I think it is necessary to detect objects even in nighttime images, but is it possible to detect objects outside of the daytime? |
Yes, sure. The current object detection algorithms can detect object in night time images. |
ICACT20220317 Question.9
Questioner: hyeonah214@skku.edu 2022-02-15 오후 10:03:13 |
ICACT20220317 Answer.9
Answer by Auhor phamdinhlam@kgu.ac.kr 2022-02-15 오후 10:03:13
Chrome Click!! |
| Thank you for the presentation. Wouldn't it be a problem with personal privacy if CCTV for real life are detected? |
Only permitted user can access to the system. For example, security man, police officers, etc. |
