Stroke Width Research Articles

Text detection in natural scene images based on color prior guided MSER

Abstract In this paper, we focus on text detection in natural scene images which is conducive to content-based wild image analysis and understanding. This task is still an open problem and usually includes two key issues: text candidate extraction and verification. For text candidate extraction, we introduce a color prior to guide the character candidate extraction by Maximally Stable Extremal Region (MSER). The principle of color prior acquirement is to obtain stroke-like textures with modified Stroke Width Transform (SWT), which is based on segmented edges. For text verification, the ideology of deep learning is adopted to distinguish text/non-text candidates. To improve classification accuracy, the results of specific task CNNs are fused. The proposed framework is evaluated on the ICDAR 2013 Robust Reading Competition database. It achieves F-score at 85.87%, which are superior over several state-of-the-art text detection methods.

Artificial Urdu Text Detection and Localization from Individual Video Frames

In current era of technology, information acquisition from images and videos become most important task due to the rapid development of data mining and machine learning.The information can be either textual, visual, or combination of these. Text appearing in images or videos is a significant source of information and plays a vital role to perceive it. Developing a unified method to detect the text is hard, as textual properties (i.e. font, size, color, illumination, orientation, etc.) may vary with the complex background. So far, multimedia and computer vision community unable yet to standardize any ideal approach to extract the text smoothly. In this paper, a novel method is proposed to detect and localize artificial Urdu text in individual video frames. Firstly, Sobel and Canny edge detection operators are applied to input frame and are merged with MSER (Maximally Stable Extremal Region) detected regions. Next, geometric constraints are applied to eliminate obvious non-text regions with large and small variations. Further refining of non-text regions is achieved by stroke width transform. SVM (Support Vector Machine) classifier is trained to classify text and non-text objects. Finally, bounding boxes are used to localize the text.Experimental results show that the proposed method is robust and efficient than state-of-the-art methods.

Text Separation from Graphics by Analyzing Stroke Width Variety in Persian City Maps

Text segmentation is a live research field with vast new areas to be explored. Separating text layer from graphics is a fundamental step to exploit text and graphics information. The language used in the map is a challenging issue in text layer separation problem. All current methods are proposed for non-Persian language maps. In Persian, text strings are composed of one or more subwords. Each subword is also composed of one to several letters connected together. Therefore, the components of the text strings in Persian are more diverse in terms of size and geometric form than in English. Thus, the overlapping of the Persian text and the lines usually produces a complex structure that the existing methods cannot handle with the necessary efficiency. For this purpose, the stroke width variety of the input map is calculated, and then the average line width of graphics is estimated by analyzing the content of stroke width. After finding the average width of graphical lines, we classify the complex structure into text and graphics in pixel level. We evaluate our method on some variety of full crossing text and graphics in Persian maps and show that some promising results in terms of precision and recall (above 80% and 90%, respectively) are obtained.

Visual substitution system for room labels identification based on text detection and recognition

Vision is an essential human sense which is playing the most important role in human perception about surrounding environment. However such information is generally inaccessible for blind people. In order to enable these persons to overcome such incapability, we propose a system for identifying the neighbouring scene. In fact, the environment is labelled as a first step and then we present a personalised and adaptive method for labels detection and recognition. The automatic analysis involves the detection of personalised markers employed to extract the region of interest followed by the perspectives rectification based on homography estimation. We use the stroke width transform (SWT) to extract the text from the detected label and crop it into words and letters. The extracted text is identified using images correlation. Finally, as a substitution for visual modality, a vocal transformation of the recognised text is performed.

Degraded document image binarization using structural symmetry of strokes

This paper presents an effective approach for the local threshold binarization of degraded document images. We utilize the structural symmetric pixels (SSPs) to calculate the local threshold in neighborhood and the voting result of multiple thresholds will determine whether one pixel belongs to the foreground or not. The SSPs are defined as the pixels around strokes whose gradient magnitudes are large enough and orientations are symmetric opposite. The compensated gradient map is used to extract the SSP so as to weaken the influence of document degradations. To extract SSP candidates with large magnitudes and distinguish the faint characters and bleed-through background, we propose an adaptive global threshold selection algorithm. To further extract pixels with opposite orientations, an iterative stroke width estimation algorithm is applied to ensure the proper size of neighborhood used in orientation judgement. At last, we present a multiple threshold vote based framework to deal with some inaccurate detections of SSP. The experimental results on seven public document image binarization datasets show that our method is accurate and robust compared with many traditional and state-of-the-art document binarization approaches based on multiple evaluation measures.

Scene text detection via extremal region based double threshold convolutional network classification.

In this paper, we present a robust text detection approach in natural images which is based on region proposal mechanism. A powerful low-level detector named saliency enhanced-MSER extended from the widely-used MSER is proposed by incorporating saliency detection methods, which ensures a high recall rate. Given a natural image, character candidates are extracted from three channels in a perception-based illumination invariant color space by saliency-enhanced MSER algorithm. A discriminative convolutional neural network (CNN) is jointly trained with multi-level information including pixel-level and character-level information as character candidate classifier. Each image patch is classified as strong text, weak text and non-text by double threshold filtering instead of conventional one-step classification, leveraging confident scores obtained via CNN. To further prune non-text regions, we develop a recursive neighborhood search algorithm to track credible texts from weak text set. Finally, characters are grouped into text lines using heuristic features such as spatial location, size, color, and stroke width. We compare our approach with several state-of-the-art methods, and experiments show that our method achieves competitive performance on public datasets ICDAR 2011 and ICDAR 2013.

Deteksi Teks Secara Otomatis Pada Natural Image Berbasis Superpixel Menggunakan Maximally Stable Extremal Regions dan Stroke Width Transform

Text detection in natural image is something to do before performing character recognition. The process of text detection plays an important role in the acquisition of information in an image. This research aims to detect text automatically in natural image based on superpixels with Maximally Stable Extremal Regions (MSER) and Stroke Width Transform (SWT). The superpixel method used is Simple Linear Iterative Clustering (SLIC). The SLIC method is used for segmenting text images into superpixel spaces. Image segmentation to superpixel aims to group pixels into homogeneous regions that capture redundant images. SLIC is a technique that effectively divides images into homogeneous regions (superpixels). Furthermore MSER is used as a feature to locate the text candidate region in a segmented image with superpixel. Then edge detection is done to validate the text area that has been found. Next, the SWT method is used to distinguish both text and non-text image regions. The dataset used is ICDAR 2003. Based on test result, MSER with superpixel is able to detect region of text in natural image. SWT is also able to recover the region which is the candidate of the text in natural image.

Extracting Maya Glyphs from Degraded Ancient Documents via Image Segmentation

We present a system for automatically extracting hieroglyph strokes from images of degraded ancient Maya codices. Our system adopts a region-based image segmentation framework. Multi-resolution super-pixels are first extracted to represent each image. A Support Vector Machine (SVM) classifier is used to label each super-pixel region with a probability to belong to foreground glyph strokes. Pixelwise probability maps from multiple super-pixel resolution scales are then aggregated to cope with various stroke widths and background noise. A fully connected Conditional Random Field model is then applied to improve the labeling consistency. Segmentation results show that our system preserves delicate local details of the historic Maya glyphs with various stroke widths and also reduces background noise. As an application, we conduct retrieval experiments using the extracted binary images. Experimental results show that our automatically extracted glyph strokes achieve comparable retrieval results to those obtained using glyphs manually segmented by epigraphers in our team.

On Detecting Road Regions in a Single UAV Image

Automatic detection of road regions in aerial images remains a challenging research topic. Most existing approaches work well on the requirement of users to provide some seedlike points/strokes in the road area as the initial location of road regions, or detecting particular roads such as well-paved roads or straight roads. This paper presents a fully automatic approach that can detect generic roads from a single unmanned aerial vehicles (UAV) image. The proposed method consists of two major components: automatic generation of road/nonroad seeds and seeded segmentation of road areas. To know where roads probably are (i.e., road seeds), a distinct road feature is proposed based on the stroke width transformation (SWT) of road image. To the best of our knowledge, it is the first time to introduce SWT as road features, which show the effectiveness on capturing road areas in images in our experiments. Different road features, including the SWT-based geometry information, colors, and width, are then combined to classify road candidates. Based on the candidates, a Gaussian mixture model is built to produce road seeds and background seeds. Finally, starting from these road and background seeds, a convex active contour model segmentation is proposed to extract whole road regions. Experimental results on varieties of UAV images demonstrate the effectiveness of the proposed method. Comparison with existing techniques shows the robustness and accuracy of our method to different roads.

Script independent approach for multi-oriented text detection in scene image

Developing a text detection method which is invariant to scripts in natural scene images is a challenging task due to different geometrical structures of various scripts. Besides, multi-oriented of text lines in natural scene images make the problem more challenging. This paper proposes to explore ring radius transform (RRT) for text detection in multi-oriented and multi-script environments. The method finds component regions based on convex hull to generate radius matrices using RRT. It is a fact that RRT provides low radius values for the pixels that are near to edges, constant radius values for the pixels that represent stroke width, and high radius values that represent holes created in background and convex hull because of the regular structures of text components. We apply k-means clustering on the radius matrices to group such spatially coherent regions into individual clusters. Then the proposed method studies the radius values of such cluster components that are close to the centroid and far from the centroid to detect text components. Furthermore, we have developed a Bangla dataset (named as ISI-UM dataset) and propose a semi-automatic system for generating its ground truth for text detection of arbitrary orientations, which can be used by the researchers for text detection and recognition in the future. The ground truth will be released to public. Experimental results on our ISI-UM data and other standard datasets, namely, ICDAR 2013 scene, SVT and MSRA data, show that the proposed method outperforms the existing methods in terms of multi-lingual and multi-oriented text detection ability.

A Shape-Aware Road Detection Method for Aerial Images

Road detection in aerial images is a crucial technique for visual navigation and scene understanding in relation to unmanned aerial vehicles (UAVs). A shape-aware road detection method for aerial images is proposed in this paper. It first employs the stroke width transform (SWT) and a geodesic distance based superpixel clustering to generate proposal regions. Then, a shape classification is responsible for selecting all potential road segments from the proposal regions which appear to be long and with consistent width. All road segments selected are clustered into several groups based on width and color features. A global graph based labeling model is then applied based on each group to remove potential background clutters, as well as to generate the final output. Experiments on two public datasets demonstrate that the proposed method can handle more diverse and challenging road scenes and needs less pre-training, leading to better performance compared to conventional methods.

Integrated Method for Text Detection in Natural Scene Images

In this paper, we present a novel image operator to extract textual information in natural scene images. First, a powerful refiner called the Stroke Color Extension, which extends the widely used Stroke Width Transform by incorporating color information of strokes, is proposed to achieve significantly enhanced performance on intra-character connection and non-character removal. Second, a character classifier is trained by using gradient features. The classifier not only eliminates non-character components but also remains a large number of characters. Third, an effective extractor called the Character Color Transform combines color information of characters and geometry features. It is used to extract potential characters which are not correctly extracted in previous steps. Fourth, a Convolutional Neural Network model is used to verify text candidates, improving the performance of text detection. The proposed technique is tested on two public datasets, i.e., ICDAR2011 dataset and ICDAR2013 dataset. The experimental results show that our approach achieves state-of-the-art performance.

System recognizing Bahamian license plate with touching characters

Various methods are proposed for license plate recognition, but none of them are universal. Some common methods for license plate localization, character extraction, and recognition are analyzed. Then a system is proposed to recognize the Bahamian license plate with touching characters. A vertical edge-based method with a modified sliding window technique is used to locate the license plate, and a machine learning process is used to trim the region. The located license plate is rectified by using the minimum enclosing box and the stroke width value. Then the vertical projection and pairs of extreme points are combined to segment the characters. Finally, a deep learning method is used to recognize the characters. 2996 images are experimented on and the total recognition accuracy achieves 83.29%. Typical methods of each stage are implemented to compare with the proposed methods. In addition, the proposed system is experimented on a public dataset to show the generalization ability of the system. The experimental results show that the proposed system performs better than the other methods and is able to be used in a real-time application.

PROTOCOL OF FACTORS ASSESSMENT INFLUENTIAL IN SUSCEPTIBILITY TO EROSION OF RIVER SLOPES

ABSTRACT Actions to conserve rivers must be preceded by an environmental diagnosis, however, beyond the body of water, the adjacent atmosphere must also be characterized by ability of recognizing the human impacts and differentiate them of the natural variation of these ecosystems. This study aimed to develop a rapid assessment protocol (PAR), composed of seven parameters - vegetation, soil texture, part of the river, slope inclination slope gradient, soil depth in slope, stroke width of water course, use and land cover -, assigning them weights 0-4, in order to assess the influence of their categories in susceptibility to erosion of the river slope. The PAR was applied in 40 sampling unities (UA) in the watershed of Arroio Val de Buia in Silveira Martins, RS. From the final result of the PAR it was created erosion susceptibility classes, "stable", "susceptible" and "instable". The t test showed significant differences between the evaluated units and the Hierarchical Cluster Analysis (AHA) distinguished two classes of UA, according to the categories of susceptibility to erosion resulting from PAR. The basic parameters that determined the groupings were: soil, soil depth in slope, vegetation, the stretch of water course and, land use and land cover. We conclude that the PAR is presented as a good rapid assessment tool river, similar to the studied stream, being useful to the environmental planning considering the regional characteristics.

An enhanced text detection technique for the visually impaired to read text

An enhanced text detection technique (ETDT) is proposed, which is expected to aid the visually impaired to overcome their reading challenges. This work enhances the edge-preserving maximally stable extremal regions (eMSER) algorithm using the pyramid histogram of oriented gradients (PHOG). Histogram of oriented gradients (HOG) derived from different pyramid levels is important while detecting maximally stable extremal regions (MSER) in the ETDT approach because it gives more spatial information when compared to HOG information from a single level. To group text, a four-line, text-grouping method is newly designed for this work. Also, a new text feature, Shapeness Score is proposed, which significantly identifies text regions when combined with the other features based on morphology and stroke widths. Using the feature vector of dimension 10, the J48 decision tree and AdaBoost machine learning algorithms identify the text regions in the images. The algorithm yields better results than the existing benchmark algorithms for the ICDAR 2011 born-digital dataset and must be improved with respect to the scene text dataset.

Arbitrarily-oriented multi-lingual text detection in video

Text detection in arbitrarily-oriented multi-lingual video is an emerging area of research because it plays a vital role for developing real-time indexing and retrieval systems. In this paper, we propose to explore moments for identifying text candidates. We introduce a novel idea for determining automatic windows to extract moments for tackling multi-font and multi-sized text in video based on stroke width information. The temporal information is explored to find deviations between moving and non-moving pixels in successive frames iteratively, which results in static clusters containing caption text and dynamic clusters containing scene text, as well as background pixels. The gradient directions of pixels in static and dynamic clusters are analyzed to identify the potential text candidates. Furthermore, boundary growing is proposed that expands the boundary of potential text candidates until it finds neighbor components based on the nearest neighbor criterion. This process outputs text lines appearing in the video. Experimental results on standard video data, namely, ICDAR 2013, ICDAR 2015, YVT videos and on our own English and Multi-lingual videos demonstrate that the proposed method outperforms the state-of-the-art methods.

Effects of flanker type and position on foveal letter recognition

Our ability to identify a foveally viewed letter at resolution threshold reduces when flankers are placed at a certain distance from it, compared to when it is presented on its own. In this study, we investigated how type and position of flankers influences foveal letter recognition. We measured participants’ performance to identify unflanked Sheridan-Gardiner letters (A, H, O, U, T, V or X), using a seven-alternative-forced-choice paradigm with the method of constant stimuli to obtain 80-90% correct response. Performance was measured again to identify a target letter in the presence of different flanker types. Flankers were bars and letters (placed either to the left and right of the target, above and below the target or on four sides of the target), and a surrounding box. Separation between the target letter and flankers varied between zero (abutting) and ten stroke widths (two letter sizes). For all flanker types, separation between the target ad the flankers significantly influenced performance (all p < 0.0001). Flankers abutting the target and at separation of one stroke width caused the largest reduction in performance, regardless of flanker type. For bar flankers, the largest drop in performance (up to 50 percent) occurred with bars placed on all four sides of the target. For letter flankers, flankers placed above and below the target reduced performance as much as four surrounding flankers. At separation of one stroke width, flanking letters and bars caused a further 10% reduction in performance than a surrounding box. Our results would have significant implications on the design of crowded visual acuity charts, especially those that are available on the market, which have different crowding features between them.

Effects of Font Size, Stroke Width, and Character Complexity on the Legibility of Chinese Characters

AbstractThis study investigated the effects of font size, stroke width, and character complexity on the legibility of Chinese characters. A within‐subjects design experiment was employed. Thirty‐six subjects participated in the experiment and completed the character‐search tasks in the pseudo‐texts. The search time per target character, correct response number, and correct response rate were used to measure the legibility. The results indicated that the font size and character complexity had significant effects on the legibility, while the effect of stroke width was not significant. All the two‐way interactions among the three display factors had significant effects on the legibility. Stroke width is critical to the legibility for characters with high‐level complexity, whereas font size is the critical factor influencing the legibility for characters with low‐level complexity. The combination of 12pt font size and 1:10 stroke width resulted in the best legibility for both character complexity levels. The findings provide useful information for the font design of Chinese texts displayed on a visual display terminal to improve the legibility.

Stroke width‐based directional total variation regularisation for document image super resolution

In the Bayesian image super resolution (SR), a regularisation term is minimised along with a data-fidelity term to generate a high-resolution (HR) image from input low-resolution (LR) image. The regularisation term is incorporated into the SR to fulfil a prior knowledge over the HR image. For instance, smoothness in the background and foreground regions is a prior knowledge about document images. The bilateral total variation (BTV), as a known regularisation term, uniformly smooths the image in all directions while preserving the edges. In this study, the authors present a document image SR method by introducing a new regularisation term called the stroke width-based directional total variation (SWDTV). It is a modified version of the BTV, and adaptively performs smoothing based on the local width and local direction of the character strokes. By analysing the input LR image and its interpolation, the local stroke width and local stroke direction in the output image are approximated. This information is encapsulated into the SWDTV regularisation term. By minimising the linear combination of the regularisation and data-fidelity terms, the HR image is reconstructed. The authors’ method achieves significant improvements in visual image quality and optical character recognition accuracy compared to related non-example-based document image SR methods.

Evolution maps and applications

Common tasks in document analysis, such as binarization, line extraction etc., are still considered difficult for highly degraded text documents. Having reliable fundamental information regarding the characters of the document, such as the distribution of character dimensions and stroke width, can significantly improve the performance of these tasks. We introduce a novel perspective of the image data which maps the evolution of connected components along the change in gray scale threshold. The maps reveal significant information about the sets of elements in the document, such as characters, noise, stains, and words. The information is further employed to improve state of the art binarization algorithm, and achieve automatically character size estimation, line extraction, stroke width estimation, and feature distribution analysis, all of which are hard tasks for highly degraded documents.