Bayesian Approach to Photo Time-Stamp Recognition

Shahab, Asif; Dengel, Andreas

doi:10.1109/icdar.2011.210

Cited by 6 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While Chen et al [3] and Fumin et al [8] perform template matching with manually created digit templates, Li [12] uses Self-Generating Neural Networks (SGNNs) to model digit appearance. Chen et al [3] make no assumptions on timestamp position, Fumin et al [8] and Li [12] limit their search to only the four image corners, while Shahab et al [26] learn location priors from a training set. Unlike these works, our goal is not to segment or read timestamps, but merely to detect their presence.…”

Section: Related Workmentioning

confidence: 99%

“…While several approaches exist for detecting and reading timestamps in photos [3,12,8,26], most approaches for detecting visible watermarks are targeted at videos [33,16,11]. Because they make strong assumptions about the appearance and position of the timestamps or water- marks, they would fail when applied to the variety of WTFs present in Internet photos.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fixing WTFs: Detecting Image Matches Caused by Watermarks, Timestamps, and Frames in Internet Photos

Weyand

Tsai

Leibe

2015

2015 IEEE Winter Conference on Applications of Computer Vision

View full text Add to dashboard Cite

An increasing number of photos in Internet photo collections comes with watermarks, timestamps, or frames (in the following called WTFs) embedded in the image content. In image retrieval, such WTFs often cause false-positive matches. In image clustering, these false-positive matches can cause clusters of different buildings to be joined into one. This harms applications like landmark recognition or large-scale structure-from-motion, which rely on clean building clusters. We propose a simple, but highly effective detector for such false-positive matches. Given a matching image pair with an estimated homography, we first determine similar regions in both images. Exploiting the fact that WTFs typically appear near the border, we build a spatial histogram of the similar regions and apply a binary classifier to decide whether the match is due to a WTF. Based on a large-scale dataset of WTFs we collected from Internet photo collections, we show that our approach is general enough to recognize a large variety of watermarks, timestamps, and frames, and that it is efficient enough for largescale applications. In addition, we show that our method fixes the problems that WTFs cause in image clustering applications. The source code is publicly available 1 and easy to integrate into existing retrieval and clustering systems.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Fixing WTFs: Detecting Image Matches Caused by Watermarks, Timestamps, and Frames in Internet Photos

Weyand

Tsai

Leibe

2015

2015 IEEE Winter Conference on Applications of Computer Vision

View full text Add to dashboard Cite

show abstract

“…In the recent years, visual attention models have been employed for various object detection/recognition tasks [8], [9], [10]. Though the usage of visual attention models for character detection is still under-investigated, their effectiveness has been shown by Shahab et al [11,12] and Uchida et al [13]. Those researchers, who try to employ visual attention models for scene character detection, believe that the characters have different properties compared with their non-character neighbors (pop-out).…”

Section: Introductionmentioning

confidence: 99%

A Hierarchical Visual Saliency Model for Character Detection in Natural Scenes

Gao

Uchida

Feng

2014

Camera-Based Document Analysis and Recognition

Self Cite

View full text Add to dashboard Cite

Abstract. Visual saliency models have been introduced to the field of character recognition for detecting characters in natural scenes. Researchers believe that characters have different visual properties from their non-character neighbors, which make them salient. With this assumption, characters should response well to computational models of visual saliency. However in some situations, characters belonging to scene text mignt not be as salient as one might expect. For instance, a signboard is usually very salient but the characters on the signboard might not necessarily be so salient globally. In order to analyze this hypothesis in more depth, we first give a view of how much these background regions, such as sign boards, affect the task of saliency-based character detection in natural scenes. Then we propose a hierarchical-saliency method for detecting characters in natural scenes. Experiments on a dataset with over 3,000 images containing scene text show that when using saliency alone for scene text detection, our proposed hierarchical method is able to capture a larger percentage of text pixels as compared to the conventional single-pass algorithm.

show abstract

Text Localization and Recognition in Images and Video

Uchida¹

2014

Handbook of Document Image Processing and Recognition

View full text Add to dashboard Cite

Bayesian Approach to Photo Time-Stamp Recognition

Cited by 6 publications

References 12 publications

Fixing WTFs: Detecting Image Matches Caused by Watermarks, Timestamps, and Frames in Internet Photos

Fixing WTFs: Detecting Image Matches Caused by Watermarks, Timestamps, and Frames in Internet Photos

A Hierarchical Visual Saliency Model for Character Detection in Natural Scenes

Text Localization and Recognition in Images and Video

Contact Info

Product

Resources

About