Pixel Aggregation Network

This is an unofficial TensorFlow re-implementation of paper "Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network" published in ICCV 2019, with Tensorflow 2.

Task

• [x] Backbone model
• [x] FPEM model
• [x] FFM model
• [x] Integrated model
• [x] Loss Function
• [x] Data preprocessing
• [x] Data postprocessing
• [x] Training pipeline
• [x] Inference pipeline
• [x] Evaluation pipeline

Command

Training

python train.py --batch 16 --epoch 601 --dataset_type ctw --gpu True

Inference

python inference.py --input ./data/CTW1500/test/text_image --model ./outputs/model_epoch_600 --bbox_type poly

Results

*Default using ResNet-50 backbone model, also support ResNet-101 backbone model.

CTW1500

Supported Dataset

• [x] CTW1500: https://github.com/Yuliang-Liu/Curve-Text-Detector

Source

[1] Original paper: https://arxiv.org/abs/1908.05900

[2] Official PyTorch code: https://github.com/whai362/pan_pp.pytorch