GA-DAN models cross-domain shifts in geometry and appearance spaces with multi-modal spatial learning and disentangled cycle-consistency loss, yielding superior scene text detection and recognition performance on adapted images.
Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
abstract
In this work we present a framework for the recognition of natural scene text. Our framework does not require any human-labelled data, and performs word recognition on the whole image holistically, departing from the character based recognition systems of the past. The deep neural network models at the centre of this framework are trained solely on data produced by a synthetic text generation engine -- synthetic data that is highly realistic and sufficient to replace real data, giving us infinite amounts of training data. This excess of data exposes new possibilities for word recognition models, and here we consider three models, each one "reading" words in a different way: via 90k-way dictionary encoding, character sequence encoding, and bag-of-N-grams encoding. In the scenarios of language based and completely unconstrained text recognition we greatly improve upon state-of-the-art performance on standard datasets, using our fast, simple machinery and requiring zero data-acquisition costs.
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
BLS approximates per-sample loss importance via EMA of batch losses, enabling simple and effective dynamic pruning of 20-50% samples losslessly across many datasets and models.
OCRBench provides the largest evaluation suite yet for OCR capabilities in large multimodal models, revealing gaps in multilingual, handwritten, and mathematical text handling.
GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data.
A dual-stream fully convolutional network produces competitive character error rates on IAM and RIMES handwriting datasets while avoiding CTC, dictionaries, and heavy preprocessing.
TTS-generated numeric training data plus a compact neural denormalizer improve E2E ASR word error rates on numeric sequences by up to a factor of 8 for the longest cases.
Survey proposing a taxonomy for document parsing into pipeline-based systems and VLM-driven unified models, reviewing components, metrics, benchmarks, and challenges.
citing papers explorer
-
GA-DAN: Geometry-Aware Domain Adaptation Network for Scene Text Detection and Recognition
GA-DAN models cross-domain shifts in geometry and appearance spaces with multi-modal spatial learning and disentangled cycle-consistency loss, yielding superior scene text detection and recognition performance on adapted images.
-
Batch Loss Score for Dynamic Data Pruning
BLS approximates per-sample loss importance via EMA of batch losses, enabling simple and effective dynamic pruning of 20-50% samples losslessly across many datasets and models.
-
OCRBench: On the Hidden Mystery of OCR in Large Multimodal Models
OCRBench provides the largest evaluation suite yet for OCR capabilities in large multimodal models, revealing gaps in multilingual, handwritten, and mathematical text handling.
-
GIT: A Generative Image-to-text Transformer for Vision and Language
GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data.
-
Fully Convolutional Networks for Handwriting Recognition
A dual-stream fully convolutional network produces competitive character error rates on IAM and RIMES handwriting datasets while avoiding CTC, dictionaries, and heavy preprocessing.
-
Improving Performance of End-to-End ASR on Numeric Sequences
TTS-generated numeric training data plus a compact neural denormalizer improve E2E ASR word error rates on numeric sequences by up to a factor of 8 for the longest cases.
-
Document Parsing Unveiled: Techniques, Challenges, and Prospects for Structured Information Extraction
Survey proposing a taxonomy for document parsing into pipeline-based systems and VLM-driven unified models, reviewing components, metrics, benchmarks, and challenges.