paddlepaddle

DBNet++

Model Description

DBNet++ is an advanced scene text detection model that combines a Differentiable Binarization (DB) module with an Adaptive Scale Fusion (ASF) mechanism. The DB module integrates binarization directly into the segmentation network, simplifying post-processing and improving accuracy. The ASF module enhances scale robustness by adaptively fusing multi-scale features. This architecture enables DBNet++ to detect text of arbitrary shapes and extreme aspect ratios efficiently, achieving state-of-the-art performance in both accuracy and speed across various text detection benchmarks.

Supported Environments

GPU	IXUCA SDK	Release
BI-V150	3.1.1	24.03

Model Preparation

Prepare Resources

Download ICDAR 2015 Dataset.

# ICDAR2015 PATH as follow:
$ ls -al /home/datasets/ICDAR2015/text_localization
total 133420
drwxr-xr-x 4 root root      179 Jul 21 15:54 .
drwxr-xr-x 3 root root       39 Jul 21 15:50 ..
drwxr-xr-x 2 root root    12288 Jul 21 15:53 ch4_test_images
-rw-r--r-- 1 root root 44359601 Jul 21 15:51 ch4_test_images.zip
-rw-r--r-- 1 root root 90667586 Jul 21 15:51 ch4_training_images.zip
drwxr-xr-x 2 root root    24576 Jul 21 15:53 icdar_c4_train_imgs
-rw-r--r-- 1 root root   468453 Jul 21 15:54 test_icdar2015_label.txt
-rw-r--r-- 1 root root  1063118 Jul 21 15:54 train_icdar2015_label.txt

# Prepare datasets
mkdir train_data pretrain_models
ln -s /path/to/icdar2015/ train_data/icdar2015

# Pretrain
wget -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/pretrained/MobileNetV3_large_x0_5_pretrained.pdparams

Install Dependencies

# Clone PaddleOCR, branch: release/2.5
git clone -b release/2.5  https://github.com/PaddlePaddle/PaddleOCR.git

# Copy PaddleOCR 2.5 patch from toolbox
yes | cp -rf ../../../../toolbox/PaddleOCR/* PaddleOCR/
cd PaddleOCR

# install requirements.
bash ../init.sh

# build PaddleOCR
python3 setup.py develop

Model Training

# run training
export FLAGS_cudnn_exhaustive_search=True
export FLAGS_cudnn_batchnorm_spatial_persistent=True
export CUDA_VISIBLE_DEVICES=0,1,2,3
python3 -m paddle.distributed.launch --gpus $CUDA_VISIBLE_DEVICES \
    tools/train.py \
    -c configs/det/det_mv3_db.yml \
    -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained

Model Results

Model	GPUs	IPS	ACC
DBNet++	BI-V100 x8	5.46 samples/s	precision: 0.9062

References

• PaddleOCR