Wovchena / text-detection-fots.pytorch

FOTS text detection branch reimplementation, hmean: 83.3%
79 stars 21 forks source link
cvpr2018 fots pytorch text-detection

FOTS: Fast Oriented Text Spotting with a Unified Network text detection branch reimplementation (PyTorch)

Train

  1. Train with SynthText for 9 epochs
    time python3 train.py --train-folder SynthText/ --batch-size 21 --batches-before-train 2

    At this point the result was Epoch 8: 100%|█████████████| 390/390 [08:28<00:00, 1.00it/s, Mean loss=0.98050].

  2. Train with ICDAR15

    Replace a data set in data_set = datasets.SynthText(args.train_folder, datasets.transform) with datasets.ICDAR2015 in train.py and run

    time python3 train.py --train-folder icdar15/ --continue-training --batch-size 21 --batches-before-train 2

    It is expected that the provided --train-folder contains unzipped ch4_training_images and ch4_training_localization_transcription_gt. To avoid saving model at each epoch, the line if True: in train.py can be replaced with if epoch > 60 and epoch % 6 == 0:

    The result was Epoch 582: 100%|█████████████| 48/48 [01:05<00:00, 1.04s/it, Mean loss=0.11290].

Learning rate schedule:

Epoch 175: reducing learning rate of group 0 to 5.0000e-04.

Epoch 264: reducing learning rate of group 0 to 2.5000e-04.

Epoch 347: reducing learning rate of group 0 to 1.2500e-04.

Epoch 412: reducing learning rate of group 0 to 6.2500e-05.

Epoch 469: reducing learning rate of group 0 to 3.1250e-05.

Epoch 525: reducing learning rate of group 0 to 1.5625e-05.

Epoch 581: reducing learning rate of group 0 to 7.8125e-06.

Test

python3 test.py --images-folder ch4_test_images/ --output-folder res/ --checkpoint epoch_582_checkpoint.pt && zip -jmq runs/u.zip res/* && python2 script.py -g=gt.zip -s=runs/u.zip

ch4_training_images and ch4_training_localization_transcription_gt are available in Task 4.4: End to End (2015 edition). script.py and ch4_test_images can be found in My Methods (Script: IoU and test set samples).

It gives Calculated!{"precision": 0.8694968553459119, "recall": 0.7987481945113144, "hmean": 0.8326223337515684, "AP": 0}.

The pretrained models are here: https://drive.google.com/open?id=1xaVshLRrMEkb9LA46IJAZhlapQr3vyY2

test.py has a commented code to visualize results.

Difference with the paper

  1. The model is different compared to what the paper describes. An explanation is in model.py.
  2. The authors of FOTS could not train on clipped words because they also have a recognition branch. The whole word is required to be present on an image to be able to be recognized correctly. This reimplementation has only detection branch and that allows to train on crops of the words.
  3. The paper suggest using some other data sets in addition. Training on SynthText is simplified in this reimplementation.