CRNN<T>
Models & Types · AiDotNet.ComputerVision.OCR.Recognition
CRNN (Convolutional Recurrent Neural Network) for text recognition.
Inheritance: ModelBase<T, Tensor<T>, Tensor<T>> → VisionTaskModelBase<T> → OCRBase<T> → CRNN<T>
Implements: IGeneratedParameterRegistrar<T>
For Beginners
CRNN combines CNNs for visual feature extraction with RNNs (specifically LSTM) for sequence modeling. It’s trained with CTC loss to handle variable-length text without requiring character-level alignment.
How It Works
Key features:
- CNN backbone for visual features
- Bidirectional LSTM for sequence modeling
- CTC (Connectionist Temporal Classification) decoding
- Handles variable-length text naturally
Reference: Shi et al., “An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition”, TPAMI 2017
Constructors
| Constructor | Summary |
|---|---|
CRNN(OCROptions<T> options) |
Creates a new CRNN text recognizer. |
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
OCRResult<T> Recognize(Tensor<T> image) |
|
Task LoadWeightsAsync(string pathOrUrl, CancellationToken cancellationToken) |
|
long GetParameterCount() |
|
void SaveWeights(string path) |
|
void SetTrainingMode(bool training) |
|
void Train(Tensor<T> input, Tensor<T> expectedOutput) |
Runs one training step with the CTC loss. |