A python package for use Textboxes : Image Text Detection Model
implemented by tensorflow, cv2
Textboxes Paper Review in Korean (My Blog) : shinjayne.github.io/textboxes
shintb
: useable textboxes python package (Source codes are in here)
svt1
: Street view Text dataset. can use with shintb.svt_data_loader.SVTDataLoader
when training Textboxes model
config.py
: (NECESSARY) configuration of model building and training with shinTB
main.py
: simple example useage of shinTB
package
shintb
python package and config.py
for building and training your own Textboxes model.svt1
gives us training / test data.config.config
and shintb
.from config import config
from shintb import graph_drawer, default_box_control, svt_data_loader, runner
GraphDrawer
,DefaultBoxControl
,SVTDataLoader
instance.graphdrawer = graph_drawer.GraphDrawer(config)
dataloader = svt_data_loader.SVTDataLoader('./svt1/train.xml', './svt1/test.xml')
dbcontrol = default_box_control.DefaultBoxControl(config, graphdrawer)
GraphDrawer
instance contains a tensorflow graph of Textboxes.
DefaultboxControl
instance contains methods and attributes which is related to default box.
SVTDataLoader
instance loads data from svt1
.
Initialize Runner
instance.
runner = runner.Runner(config, graphdrawer, dataloader, dbcontrol)
Runner
uses GraphDrawer
,DefaultBoxControl
,SVTDataLoader
instance.Runner.train()
. Every 1000 step, shintb
will save ckpt file in the directory you set in config.py
.runner.train()
Runner.test()
runner.test()
Runner.image()
.runner.image(<your_image_directory>)
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。