Dateset is not here

[ahmed-nady]

Dateset is not here

[ahmed-nady]

Hi guy, For the UMN dataset,
There are three scenes (Lawn,Indoor, Plaza), for each each you take the first 400 frames as training and the rest for testing except the Indoor scene we you the first 300 normal frames. That's it
Also you can look at the research paper at
https://www.researchgate.net/publication/327972246_REAL-TIME_ABNORMAL_EVENT_DETECTION_IN_CROWDED_SCENES

[Littlebee1]

Thank you very much for you replay.I read your paper recently.It mainlyintruduce the global anomalous evnet detection.I want to know if this code is also for local anomalous event detection.

[Littlebee1]

There is UCSD dateset in your project, It is also for UCSD dateset?

[ahmed-nady]

yeah, you can extend the global anomalous event detection to local anomalous detection by dividing the frame into cells/regions and repeat the process of global event detection on each region.
I used UCSD dataset for that.

[Littlebee1]

I'm sorry to trouble you again，these days I used this method for ucsd dateset,while the result of local anomalous event detection is not satisfied .Could you give me some advice，thanks indeed！ ^-^