I am working on a project named as video anomaly detection, and I have decided to apply one of these algorithm on my project named as AutoEncoders, RNN, and LSTM.
Kindly guide me which one among them is best for video anomaly detection and why. It will be a great favor.
Regards,
In my opinion, you should try auto-encoders first. Take a look at my recent work based on auto-emcoders: https://www.computer.org/csdl/journal/tp/5555/01/09410375/1sYYrP4z1a8
استاذي الكريم تستطيع استخدام خوارزميات التشفير وتوضيفها لاكتشاف الشذوذ اتمنى لك النجاح والتوفيق
In most of the concepts are used LSTM for detecting video anomaly, but when comparing to LSTM Concept CNN is getting a better results.
Dear Tariq Sm
This article may be helpful:
https://www.mdpi.com/1424-8220/21/8/2811/htm
Preprint Video Trajectory Classification and Anomaly Detection Using ...
You can also use graph neural networks to capture the normal semantic of your video. See this work for more information :
Preprint Ano-Graph: Learning Normal Scene Contextual Graphs to Detect...
dear researcher better you can choose CNN for image and video related data.
Auto Encoders can be used as a base model in case of anomaly detection !
- Badges
- Science topic
- Similar topics
- Names