ASTEREncoder

class mmocr.models.textrecog.ASTEREncoder(in_channels, num_layers=2, init_cfg={'layer': 'Conv2d', 'type': 'Xavier'})

Implement BiLSTM encoder module in `ASTER: An Attentional Scene Text Recognizer with Flexible Rectification.

<https://ieeexplore.ieee.org/abstract/document/8395027/` :param in_channels: Number of input channels. :type in_channels: int :param num_layers: Layers of BiLSTM. Defaults to 2. :type num_layers: int

Parameters

• in_channels (int) –

• num_layers (int) –

Return type

None

forward(feat, img_metas=None)

Parameters

feat (Tensor) – Feature of shape (N, C, 1, W).

Returns

Output of BiLSTM.

Return type

Tensor