search

callsys / FlowText

[ICME 2023] FlowText: Synthesizing Realistic Scene Text Video with Optical Flow Estimation
11 stars 0 forks source link

Relevant instructions on training and testing #1

Open P1ng3609 opened 6 months ago

P1ng3609 commented 6 months ago

Are there instructions related to training and testing, please? I noticed that there is no relevant information in the readme. It would be great if you could provide.

callsys commented 6 months ago

The proposed data generation pipeline relies on off-the-shelf trained models and requires no training. For testing, we have released the data generation code here. To evaluate the generated data quality, you can generate some videos and use them to train existing video text spotters.