Adapting off-the-shelf CNNs for word spotting &amp;amp; recognition

Sharma, Arjun; Sankar, K. Pramod

doi:10.1109/icdar.2015.7333909

Cited by 32 publications

(21 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach can be likened to the Softmax CNN from [2] or the fine-tuned CNN from [11] where they learn features by formulating a multi-class classification problem. But the triplet-CNN (70.81% MAP on IAM) significantly outperforms both the Softmax CNN (48.67% MAP) and the fine-tuned CNN (46.53% MAP).…”

Section: Discussionmentioning

confidence: 99%

“…For the handwritten word spotting community, hand-crafted image representations, primarily Fisher vectors and other bag-of-visual-words models, have been the features of choice [1], [3], [10]. However, there have been some notable exceptions [2], [11]. A hand-crafted approach has also been used for the word embeddings, where the embedding of choice has been the PHOC representation and its variations.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Semantic and Verbatim Word Spotting Using Deep Neural Networks

Wilkinson

Brun

2016

2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR)

View full text Add to dashboard Cite

Abstract-In the last few years, deep convolutional neural networks have become ubiquitous in computer vision, achieving state-of-the-art results on problems like object detection, semantic segmentation, and image captioning. However, theyhave not yet been widely investigated in the document analysis community. In this paper, we present a word spotting system based on convolutional neural networks. We train a network to extract a powerful image representation, which we then embed into a word embedding space. This allows us to perform word spotting using both query-by-string and query-by-example in a variety of word embedding spaces, both learned and handcrafted, for verbatim as well as semantic word spotting. Our novel approach is versatile and the evaluation shows that it outperforms the previous state-of-the-art for word spotting on standard datasets.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Semantic and Verbatim Word Spotting Using Deep Neural Networks

Wilkinson

Brun

2016

2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR)

View full text Add to dashboard Cite

show abstract

“…This competition has been ruled by CNNs ever since with the winning teams always featuring "very deep" architectures [14], [15] Despite their large success, there has been very limited work on using CNNs for word spotting. In [1] a pretrained deep CNN is finetuned to learn classes of word images. The output is then used to perform word spotting.…”

Section: B Convolutional Neural Networkmentioning

confidence: 99%

“…Word spotting is an effective paradigm to index document images for which a direct classification approach would be infeasable. In [1] the authors use a pretrained CNN to perform word spotting on the IAM database. However, this approach has several short comings: Each word image has to be cropped to a unit width and height which almost always distorts the image.…”

Section: Introductionmentioning

confidence: 99%

PHOCNet: A Deep Convolutional Neural Network for Word Spotting in Handwritten Documents

Sudholt

Fink

2016

2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR)

197

204

View full text Add to dashboard Cite

In recent years, deep convolutional neural networks have achieved state of the art performance in various computer vision task such as classification, detection or segmentation. Due to their outstanding performance, CNNs are more and more used in the field of document image analysis as well. In this work, we present a CNN architecture that is trained with the recently proposed PHOC representation. We show empirically that our CNN architecture is able to outperform state of the art results for various word spotting benchmarks while exhibiting short training and test times.

show abstract

“…The first version contains binarized word images which have been slant-corrected 2 . The second version 3 contains the plain gray-level document images and is by far the one more commonly used for evaluating word spotting methods, e.g., in [3,44,[46][47][48]51]. For our experiments, we will make use of the plain gray-level document images as well.…”

Section: George Washingtonmentioning

confidence: 99%

Attribute CNNs for word spotting in handwritten documents

Sudholt

Fink

2018

IJDAR

View full text Add to dashboard Cite

Word spotting has become a field of strong research interest in document image analysis over the last years. Recently, AttributeSVMs were proposed which predict a binary attribute representation [3]. At their time, this influential method defined the state-of-theart in segmentation-based word spotting. In this work, we present an approach for learning attribute representations with Convolutional Neural Networks (CNNs). By taking a probabilistic perspective on training CNNs, we derive two different loss functions for binary and real-valued word string embeddings. In addition, we propose two different CNN architectures, specifically designed for word spotting. These architectures are able to be trained in an end-to-end fashion. In a number of experiments, we investigate the influence of different word string embeddings and optimization strategies. We show our Attribute CNNs to achieve state-of-the-art results for segmentation-based word spotting on a large variety of data sets.

show abstract

Adapting off-the-shelf CNNs for word spotting & recognition

Cited by 32 publications

References 12 publications

Semantic and Verbatim Word Spotting Using Deep Neural Networks

Semantic and Verbatim Word Spotting Using Deep Neural Networks

PHOCNet: A Deep Convolutional Neural Network for Word Spotting in Handwritten Documents

Attribute CNNs for word spotting in handwritten documents

Contact Info

Product

Resources

About