search

coranholmes / TEVAD

Official implementation for paper TEVAD: Improved video anomaly detection with captions
23 stars 5 forks source link

Results in Figure 3. #4

Closed MoshiraTolbah closed 1 year ago

MoshiraTolbah commented 1 year ago

Hello, Can you please tell me how you produced the results shown in your paper in Figure 3? (Example results from (a) ShanghaiTech (riding a bike), (b) XD-Violence (riot) , and (c) UCF-Crime (vandalism) datasets.)

coranholmes commented 1 year ago

Hi, the visualization codes are a bit messy so I am afraid it's hard for me to share them with you. With other codes (as well as the captions generated) provided in this repo, you should be able to generate similar figures easily.

coranholmes commented 1 year ago

Unfortunately, I have left the organization where I did this research so I can't find the source codes myself. The general steps are: 1) generate the line charts (matplotlib) 2) get the screen shots 3) use Powerpoint to combine the line chart, the screen shot and the captions.