H
Huaigu Cao
Researcher at Raytheon
Publications - 49
Citations - 594
Huaigu Cao is an academic researcher from Raytheon. The author has contributed to research in topics: Optical character recognition & Handwriting recognition. The author has an hindex of 15, co-authored 49 publications receiving 538 citations. Previous affiliations of Huaigu Cao include Northwestern University & BBN Technologies.
Papers
More filters
Proceedings ArticleDOI
Gabor features for offline Arabic handwriting recognition
TL;DR: This work proposes Gabor features, a set of binarization based features which have been proven to be effective in capturing shape characteristics of handwritten Arabic subwords and are likely to be more robust to noises in document images.
Proceedings ArticleDOI
Automated image quality assessment for camera-captured OCR
TL;DR: A novel automated image quality assessment method that predicts the degree of degradation on OCR, and quantifies image quality degradation across several artifacts and accurately predicts the impact on O CR error rate is proposed.
Proceedings ArticleDOI
Improvements in BBN's HMM-Based Offline Arabic Handwriting Recognition System
TL;DR: A novel integration of structural features in the HMM framework which exclusively results in a 9% relative improvement in performance is proposed, and a relative reduction of 17% in word error rate over the baseline Arabic handwriting recognition system is demonstrated.
Proceedings ArticleDOI
Using Convolutional Encoder-Decoder for Document Image Binarization
TL;DR: The proposed convolutional encoder-decoder model with deep learning for document image binarization has comparable performance to the other hand-crafted binarized approaches and has more generalization capabilities with limited in-domain training data.
Proceedings ArticleDOI
Combining Convolutional Neural Networks and LSTMs for Segmentation-Free OCR
TL;DR: A novel end-to-end trainable OCR system combining a CNN for feature extraction with 1-D LSTMs for sequence modeling that will make it easier to use techniques borrowed from CNN research in computer vision to improve OCR performance.