search

clovaai / deep-text-recognition-benchmark

Text recognition (optical character recognition) with deep learning methods, ICCV 2019
Apache License 2.0
3.71k stars 1.09k forks source link

Lexicon-50, Lexicon-100 lmdb dataset #195

Open ducMNSD opened 4 years ago

ducMNSD commented 4 years ago

Could you create Lexicon-50, Lexicon-100 lmdb dataset for evalution ? Thank you very much

ku21fan commented 4 years ago

Hello,

I made and tested them before, but I don't have them right now...

You can create them with create_lmdb_dataset.py.

Just read lexicon files and save them as follows.

        imageKey = 'image-%09d'.encode() % cnt
        labelKey = 'label-%09d'.encode() % cnt
        lexiconKey = 'lexicon-%09d'.encode() % cnt
        cache[imageKey] = imageBin
        cache[labelKey] = label.encode()
        cache[lexiconKey ] = lexicon.encode()

Hope it helps,

Best.