What is Twitter

Scene text detection methods based on neural networks have emerged recently and have shown promising results. Previous methods trained with rigid word-level bounding boxes exhibit limitations in representing the text region in an arbitrary shape. In this paper, we propose a new scene text detection method to effectively detect text area by exploring each character and affinity between characters. To overcome the lack of individual character level annotations, our proposed framework exploits both the given character-level annotations for synthetic images and the estimated character-level ground-truths for real images acquired by the learned interim model. In order to estimate affinity between characters, the network is trained with the newly proposed representation for affinity. Extensive experiments on six benchmarks, including the TotalText and CTW-1500 datasets which contain highly curved texts in natural images, demonstrate that our character-level text detection significantly outperforms the state-of-the-art detectors. According to the results, our proposed method guarantees high flexibility in detecting complicated scene text images, such as arbitrarily-oriented, curved, or deformed texts.

/pdf/character-region-awareness-for-text-detection-1eceizykl0.pdf

Character Region Awareness for Text Detection

Recently, models based on deep neural networks have dominated the fields of scene text detection and recognition. In this paper, we investigate the problem of scene text spotting, which aims at simultaneous text detection and recognition in natural images. An end-to-end trainable neural network model for scene text spotting is proposed. The proposed model, named as Mask TextSpotter, is inspired by the newly published work Mask R-CNN. Different from previous methods that also accomplish text spotting with end-to-end trainable deep neural networks, Mask TextSpotter takes advantage of simple and smooth end-to-end learning procedure, in which precise text detection and recognition are acquired via semantic segmentation. Moreover, it is superior to previous methods in handling text instances of irregular shapes, for example, curved text. Experiments on ICDAR2013, ICDAR2015 and Total-Text demonstrate that the proposed method achieves state-of-the-art results in both scene text detection and end-to-end text recognition tasks.

Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes

Rotation detection is a challenging task due to the difficulties of locating the multi-angle objects and separating them effectively from the background. Though considerable progress has been made, for practical settings, there still exist challenges for rotating objects with large aspect ratio, dense distribution and category extremely imbalance. In this paper, we propose an end-to-end refined single-stage rotation detector for fast and accurate object detection by using a progressive regression approach from coarse to fine granularity. Considering the shortcoming of feature misalignment in existing refined single-stage detector, we design a feature refinement module to improve detection performance by getting more accurate features. The key idea of feature refinement module is to re-encode the position information of the current refined bounding box to the corresponding feature points through pixel-wise feature interpolation to realize feature reconstruction and alignment. For more accurate rotation estimation, an approximate SkewIoU loss is proposed to solve the problem that the calculation of SkewIoU is not derivable. Experiments on three popular remote sensing public datasets DOTA, HRSC2016, UCAS-AOD as well as one scene text dataset ICDAR2015 show the effectiveness of our approach. Tensorflow and Pytorch version codes are available at this https URL and this https URL, and R3Det is also integrated in our open source rotation detection benchmark: this https URL.

R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object

Arbitrary-oriented object detection has recently attracted increasing attention in vision for their importance in aerial imagery, scene text, and face etc. In this paper, we show that existing regression-based rotation detectors suffer the problem of discontinuous boundaries, which is directly caused by angular periodicity or corner ordering. By a careful study, we find the root cause is that the ideal predictions are beyond the defined range. We design a new rotation detection baseline, to address the boundary problem by transforming angular prediction from a regression problem to a classification task with little accuracy loss, whereby high-precision angle classification is devised in contrast to previous works using coarse-granularity in rotation detection. We also propose a circular smooth label (CSL) technique to handle the periodicity of the angle and increase the error tolerance to adjacent angles. We further introduce four window functions in CSL and explore the effect of different window radius sizes on detection performance. Extensive experiments and visual analysis on two large-scale public datasets for aerial images i.e. DOTA, HRSC2016, as well as scene text dataset ICDAR2015 and MLT, show the effectiveness of our approach. The code is public available at https://github.com/Thinklab-SJTU/CSL_RetinaNet_Tensorflow.

Arbitrary-Oriented Object Detection with Circular Smooth Label

Text detection and recognition in a natural environment are key components of many applications, ranging from business card digitization to shop indexation in a street. This competition aims at assessing the ability of state-of-the-art methods to detect Multi-Lingual Text (MLT) in scene images, such as in contents gathered from the Internet media and in modern cities where multiple cultures live and communicate together. This competition is an extension of the Robust Reading Competition (RRC) which has been held since 2003 both in ICDAR and in an online context. The proposed competition is presented as a new challenge of the RRC. The dataset built for this challenge largely extends the previous RRC editions in many aspects: the multi-lingual text, the size of the dataset, the multi-oriented text, the wide variety of scenes. The dataset is comprised of 18,000 images which contain text belonging to 9 languages. The challenge is comprised of three tasks related to text detection and script classification. We have received a total of 16 participations from the research and industrial communities. This paper presents the dataset, the tasks and the findings of this RRC-MLT challenge.

ICDAR2017 Robust Reading Challenge on Multi-Lingual Scene Text Detection and Script Identification - RRC-MLT

Microblogs such as Twitter are characterized by the richness and recency of information shared by their users during major events. However, it is very challenging to automatically mine for information or for users sharing certain information due to the huge variety of unstructured stream of data shared in such microblogs. This work proposes a ranking and classification model for identifying users sharing useful information during a specified event. The model is based on a novel set of features that can be computed in real time. These features are designed such that they take into account both the on and off-topic activities of a user. Once users are characterized by a feature vector, supervised machine learning tool is trained to classify users as either prominent or not. Our model has been tested on data shared during a flooding disaster event and performed very well. The achieved results show the effectiveness of the proposed model for both the classification and ranking of prominent users in such events, and also the importance of the adjustment of the on-topic features by the off-topic ones when describing users' activities.

/pdf/prominent-users-detection-during-specific-events-by-learning-2y28227l8o.pdf

Prominent Users Detection during Specific Events by Learning On- and Off-topic Features of User Activities

Microblogs have proved their potential to attract people from all over the world to express voluntarily what is happening around them during unexpected events. However, retrieving relevant information from the huge amount of data shared in real time in these microblogs remain complex. This paper proposes a new system named MASIR for real-time information retrieval from microblogs during unexpected events. MASIR is based on a decentralized and collaborative multi-agent approach analyzing the profiles of users interested in a given event in order to detect the most prominent ones that have to be tracked in real time. Real time monitoring of these users enables a direct access to valuable fresh information. Our experiments shows that MASIR simplifies the real-time detection and tracking of the most prominent users by exploring both the old and fresh information shared during the event and outperforms the standard centrality measures by using a time-sensitive ranking model.

/pdf/masir-a-multi-agent-system-for-real-time-information-59k2kg3111.pdf

MASIR: A Multi-agent System for Real-Time Information Retrieval from Microblogs During Unexpected Events

https://hal.archives-ouvertes.fr/hal-03025566/document

Detecting prominent microblog users over crisis events phases

The response phase in a disaster case is often considered to be the most critical in terms of saving lives and dealing with irreversible damage. The timely provision of geospatial information is crucial in the decision-making process. Thus, there is a need for the integration of heterogeneous spatial databases which are inherently distributed and created under different projects by various organizations. The integration of all relevant data for timely decision making is a challenging task due to syntactic, schematic and semantic heterogeneity. This paper aims to propose a framework for the integration of heterogeneous spatial databases using novel approaches, such as web services and ontologies. We focus on providing solutions for the three levels of heterogeneity, in order to be able to interrogate the content of the different databases conveniently. Based on the proposed framework, we implemented a use case using heterogeneous data belonging to La Rochelle city in France.

/pdf/integration-of-heterogeneous-spatial-databases-for-disaster-3w8lsoj4ba.pdf

Imen Bizid

Papers

ICDAR2017 Robust Reading Challenge on Multi-Lingual Scene Text Detection and Script Identification - RRC-MLT

Prominent Users Detection during Specific Events by Learning On- and Off-topic Features of User Activities

MASIR: A Multi-agent System for Real-Time Information Retrieval from Microblogs During Unexpected Events

Detecting prominent microblog users over crisis events phases

Integration of Heterogeneous Spatial Databases for Disaster Management