Use Case and High-Level Description

This is a network for text recognition scenario. It consists of VGG16-like backbone and bidirectional LSTM encoder-decoder. The network is able to recognize school marks that should have format either <digit> or <digit>.<digit> (e.g. 4 or 3.5).

Example

-> Mark2.5

Specification

Inputs

Shape: [1x1x32x64] - An input image in the format [BxCxHxW], where:

• B - batch size
• C - number of channels
• H - image height
• W - image width

Note that the source image should be tight aligned crop with detected text converted to grayscale.

Outputs

The net outputs a blob with the shape [16, 1, 13] in the format [WxBxL], where:

• W - output sequence length
• B - batch size
• L - confidence distribution across the alphabet: "0123456789._#", where # - special blank character for CTC decoding algorithm and the character ‘’_'` replaces all non-numeric symbols.

The network output can be decoded by CTC Greedy Decoder or CTC Beam Search decoder.

Legal Information