search

ankush-me / SynthText

Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Vedaldi, Andrew Zisserman, CVPR 2016.
http://www.robots.ox.ac.uk/~vgg/data/scenetext/
Apache License 2.0
2.02k stars 622 forks source link

Loading images in script #252

Closed Didier0 closed 2 years ago

Didier0 commented 2 years ago

Hi @ankush-me

Will the script work if I make one h5 file with seg, depth and original RBG images from others than the 5 in the script?

Because in RGB there are 11040 images, in depth are 8010 and in seg are 11035.

Or do I need to merge those pictures with same name?

ankush-me commented 2 years ago

Hi @Didier0 --- you don't need to create a new h5 for additional images. You can read them directly from the disk and pass to the generation script, a more fleshed out example is given here.

ankush-me commented 2 years ago

Yes, you would need to load the corresponding depth and segmentation information (based on image name).

Didier0 commented 2 years ago

Hi @Didier0 --- you don't need to create a new h5 for additional images. You can read them directly from the disk and pass to the generation script, a more fleshed out example is given here.

Problem is that, I can't extract area and label from seg image. When I load that images from file, it says me: object has no attributes attrs, something like that.

Didier0 commented 2 years ago

@ankush-me Why the number of images are not the same in seg.h5, depth.h5 and in the bg_images folder?? bg_image folder = 11040 depth.h5 = 8010 seg.h5 = 11035

ankush-me commented 2 years ago

Filter the file names using imnames.cp.