huoyijie/AdvancedEAST
huoyijie/AdvancedEAST has 1,228 GitHub stars and 371 forks. GitHot measured 0 stars in 24 hours, -1 in 7 days, and -1 in 15 days. Last checked 2026-11-22 17:23 UTC. Last push 2022-09-09.
Progression mesurée des étoiles
- in 24h
- 0
- in 7 days
- -1
- in 15 days
- -1
- in 1 month
- Historique en cours
Variations nettes sur des périodes qui se chevauchent, pas des totaux quotidiens. Les mesures absentes restent vides. Les étoiles indiquent l’intérêt, pas la qualité.
À propos
AdvancedEAST est un algorithme utilisé pour détecter du texte dans des images de scènes. Principalement basé sur EAST, il a bénéficié d’améliorations importantes qui rendent les prédictions de textes longs plus précises.https://github.com/huoyijie/raspberrypi-car
Apprentissage automatiqueREADME.md
OriginalAdvancedEAST
AdvancedEAST is an algorithm used for Scene image text detect, which is primarily based on EAST:An Efficient and Accurate Scene Text Detector, and the significant improvement was also made, which make long text predictions more accurate. If this project is helpful to you, welcome to star. And if you have any problem, please contact me.
- email:[email protected]
- website:https://huoyijie.cn
advantages
- writen in keras, easy to read and run
- base on EAST, an advanced text detect algorithm
- easy to train the model
- significant improvement was made, long text predictions more accurate.(please see 'demo results' part bellow, and pay attention to the activation image, which starts with yellow grids, and ends with green grids.)
In my experiments,
AdvancedEast has obtained much better prediction accuracy then East,
especially on long text. Since East calculates final vertexes coordinates with
weighted mean values of predicted vertexes coordinates of all pixels. It is too
difficult to predict the 2 vertexes from the other side of the quadrangle.
See East limitations picked from original paper bellow.

project files
- config file:cfg.py,control parameters
- pre-process data: preprocess.py,resize image
- label data: label.py,produce label info
- define network network.py
- define loss function losses.py
- execute training advanced_east.py and data_generator.py
- predict predict.py and nms.py
后置处理过程说明参见 后置处理(含原理图)
network arch
- AdvancedEast

网络输出说明: 输出层分别是1位score map, 是否在文本框内;2位vertex code,是否属于文本框边界像素以及是头还是尾;4位geo,是边界像素可以预测的2个顶点坐标。所有像素构成了文本框形状,然后只用边界像素去预测回归顶点坐标。边界像素定义为黄色和绿色框内部所有像素,是用所有的边界像素预测值的加权平均来预测头或尾的短边两端的两个顶点。头和尾部分边界像素分别预测2个顶点,最后得到4个顶点坐标。
- East

setup
- python 3.6.3+
- tensorflow-gpu 1.5.0+(or tensorflow 1.5.0+)
- keras 2.1.4+
- numpy 1.14.1+
- tqdm 4.19.7+
training
tianchi ICPR dataset download 链接: https://pan.baidu.com/s/1NSyc-cHKV3IwDo6qojIrKA 密码: ye9y
prepare training data:make data root dir(icpr), copy images to root dir, and copy txts to root dir, data format details could refer to 'ICPR MTWI 2018 挑战赛二:网络图像的文本检测', Link
modify config params in cfg.py, see default values.
python preprocess.py, resize image to 256256,384384,512512,640640,736*736, and train respectively could speed up training process.
python label.py
python advanced_east.py, train entrance
python predict.py -p demo/001.png, to predict
pretrain model download(use for test) 链接: https://pan.baidu.com/s/1KO7tR_MW767ggmbTjIJpuQ 密码: kpm2
demo results



- compared with east based on vgg16
As you can see, although the text area prediction is very accurate, the vertex coordinates are not accurate enough.

License
The codes are released under the MIT License.
references
CTPN:Detecting Text in Natural Image with Connectionist Text Proposal Network
Deep Matching Prior Network: Toward Tighter Multi-oriented Text Detection
网络输出说明: 输出层分别是1位score map, 是否在文本框内;2位vertex code,是否属于文本框边界像素以及是头还是尾;4位geo,是边界像素可以预测的2个顶点坐标。所有像素构成了文本框形状,然后只用边界像素去预测回归顶点坐标。边界像素定义为黄色和绿色框内部所有像素,是用所有的边界像素预测值的加权平均来预测头或尾的短边两端的两个顶点。头和尾部分边界像素分别预测2个顶点,最后得到4个顶点坐标。
后置处理过程说明参见 后置处理(含原理图)
Explorer des projets similaires
Chargement…