SSD_scene_text_detection

SSD_scene_text_detection copied to clipboard

Issue related to training

3

Hi while executing VGG_scenetext_SSD_300x300.sh file I got following error Check failed: num_priors_ * num_loc_classes_ * 4 == bottom[0]->channels() (58464 vs. 29232) Number of priors must match number of location predictions....

pranayamhatre175

@chenxinpeng Did you use the file ssd_icdar_scenetext.py to train? I notice that the solver_param ( line 319-343 ) in the ssd_icdar_scenetext.py is different from training_file/solver.prototxt , for example the test_interval,...

HelloTobe

Dataset

2

I can't find any ICDAR dataset which can be download online, even on conference official website. Could you give a link to download these dataset? Thx!

V1ns0n

Does SSD_scene-text-detection support chinese text detection?

1

albertyou2

default boxes aspect ratio

1

Hi, thanks for your work. I have a question that do you use the default boxes aspect ratios in original ssd? like [2],[2,3],[2,3],[2,3],[2,3] and turn the flip option as Ture....

xinghedyc