atlas

Research map / Detection and segmentation / CNN features

SqueezedText: A Real-Time Scene Text Recognition by Binary Convolutional Encoder-Decoder Network

AAAI 2018 · Detection and segmentation › CNN features · approach: Arbitrary-shape scene text detection · 106 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
B-CEDNet character salience detection + Bi-RNN correction with spatial info instead of CRNNscene text recognition · IIIT5k no lexiconaccuracy: 78.2 -> 86.6Helped
B-CEDNet + Bi-RNN instead of CRNNscene text recognition · SVT lexicon 50accuracy: 96.4 -> 96.1No clear effect
binarize weights and activations (ablation)SqueezedText · IIIT5k no lexiconaccuracy: 87.0 -> 86.6No clear effect
binarize weights and activations (ablation)CEDNet · IC03 character retrievalprecisionHelped
XNOR kernel instead of baseline matmul kernel (ablation)B-CEDNet · TITAN X GPUinference speedup: 5×Helped
binary encoder-decoder replacing FC layersscene text recognizer · storagenetwork size (MB): 25.2 MB -> 4.30 MBHelped

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy