Unpaired Document Image Denoising for OCR using BiLSTM enhanced CycleGAN
Bibliographic record
Abstract
Abstract The recognition performance of optical character recognition (OCR) models can be sub-optimal when document images suffer from various degradations. Supervised deep learning methods for image enhancement can generate high-quality enhanced images. However, these methods demand the availability of corresponding clean images or ground truth text. Sometimes this requirement is difficult to fulfill for real-world noisy documents. For instance, it can be challenging to create paired noisy/clean training datasets or obtain ground truth text for noisy point-of-sale receipts and invoices. Unsupervised methods have been explored in recent years to enhance images in the absence of ground truth images or text. However, these methods focus on enhancing natural scene images. In the case of document images, preserving the readability of text in the enhanced images is of utmost importance for improved OCR performance. In this work, we propose a modified architecture to the CycleGAN model to improve its performance in enhancing document images with better text preservation. Inspired by the success of CNN-BiLSTM combination networks in text recognition models, we propose modifying the discriminator network in the CycleGAN model to a combined CNN-BiLSTM network for better feature extraction from document images during classification by the discriminator network. Results indicate that our proposed model not only leads to better preservation of text and improved OCR performance over the CycleGAN model but also achieves better performance than the classical unsupervised image pre-processing techniques like Sauvola and Otsu.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".