141 results about "Stroke width" patented technology

The invention discloses a Chinese detection method in a natural scene image based on a connected domain. The method comprises the steps: firstly, obtaining the natural scene image, and carrying out preprocessing of stroke width transformation on the natural scene image, wherein an output of the stroke width transformation is an image, and each pixel value of the image represents a maximum possible stroke width of a pixel of a corresponding position of an original image; demarcating connected domains in the image pixels; extracting various features of a connected component after the connected component is obtained, wherein the feature combinations can well express the connected component; from a Chinese structure, firstly, carrying out within character mergence, then carrying out inter-character mergence, wherein a within character mergence method is used for detecting single Chinese character, a inter-character mergence method is used for detecting text lines, and a text area is demarcated by a rectangular frame. According to the Chinese detection method, a complex structure of Chinese serves as a starting point, pertinence in a Chinese detection aspect is good, and therefore the Chinese detection method in the natural scene image based on the connected domain has high initiative and high accuracy.

Apparatus and methods for dynamically rendering digital ink strokes of arbitrary stroke width while a user is drawing the ink strokes. An ink rendering system may receive sampled pen tip positions associated with a stylus or pen on a digitizer. The ink rendering system may further determine pen tip instances having particular shapes, sizes, and / or rotations, and may determine connecting shapes, preferably quadrangles, that connect between the pen tip instances. Smoothing techniques may also be used for existing digital ink.

The present invention discloses a rapid image text detection method based on a multi-channel and multi-dimensional cascade filter. The problem is mainly solved that the recall ratio is low and the speed is slow in the prior art. The method comprises: 1) extracting a maximum stable extremum region in the different channels and scales of an input image as a character candidate region; 2) removing the background region in the character candidate region by employing a cascade filter from coarse to fine, namely setting a threshold value for the morphological features of the character candidate region, and performing the first grade coarse filtration; setting thresholds for the stroke width and the stroke width variable coefficient of the character candidate region, performing the second grade coarse filtration, then removing the overlapping regions, and employing a convolution neural network binary classifier perform fine filtration; and 3) aggregating the region into the character string according to the geometry and the position feature of the character candidate region after cascade filter through a graph model. The rapid image text detection method based on the multi-channel and multi-dimensional cascade filter has high recall ratio, high accuracy and fast speed, and can be used for detection of image text at various interference surroundings.

The invention provides a method for recognizing characters of a licence plate based on a Haar-like feature vector and a support vector machine (SVM), belonging to the fields of pattern recognition and intelligent transportation, and relating to character image feature extraction and character classifier training. The method for recognizing the characters of the licence plate is the core technology of licence plate recognition, wherein two key problems to be solved are character feature extraction and design of a character classifier. In the invention, a Haar-like feature structure appropriate for the stroke width of a character is selected to describe a stoke of the character, and the Haar-like feature structure membership degree of a character image block is extracted to construct a feature vector for character recognition so as to train a SVM character classifier with good generalization performance. The geometric structure of the character is a key feature for the character recognition, and the geometric structure of the stroke is converted into a statistic value to describe the stroke, which is innovation of the invention. The method for recognizing the character image of the licence plate provided by the invention has good anti-interference performance, can be used for recognition of the characters of the licence plate in a traffic video and has the characteristics of good instantaneity and high recognition accuracy.

The invention relates to a digitization method and a software system capable of demonstrating a word writing process. Based on digital information, in the whole word writing demonstration process from programming of word shape basic control information to calculation of word display control information, to forming of word writing demonstration information and to final displaying of corresponding word shapes on an intelligent device, a demonstrated word can be zoomed, rotated and inclined, the color of the demonstrated word can be adjusted, the writing demonstration time can be controlled, the stroke width can be adjusted, single-stroke writing demonstration can be performed, a digital copybook can be made and other relevant aims can be fulfilled. As for writing demonstration of letters and complicated words and aiming at application of intelligent devices having different performance types, detail instructions are made according to tip word demonstration and line word demonstration. Digitalization of language word writing education is realized, and popularization and normalization of education in schools and families for language word writing as well as outside mutual communication are facilitated. The writing demonstration of Chinese characters on a computer is instructed in detail.

The invention provides a Harris angular point and stroke width based text region detection method. The method mainly comprises such three parts of edge enhanced Harris angular point detection, stroke width based candidate region scanning and text color based region expansion. According to the invention, a text region is acquired through detection for angular points, connected region analysis is carried out, and a response region which does not conform to text line characteristics is removed, thereby being capable of significantly improving the accuracy of angular point detection for the text region in an image. Compared with a text detection method singly based on edge characteristics, the method provided by the invention can improve the recall rate of text detection in the image. Compared with a text region detection method singly based on connected regions, the method provided by the invention can acquire higher detection efficiency. The text region detection method has advantages when being compared with the text detection method singly based on the edge characteristics and the text detection method singly based on the connected regions.

The invention discloses a video Optical Character Recognition (OCR) image-text separation method and a system thereof to solve the problem that the current recognition method by using a bilateral model can not extract intact strokes in the case of uneven thickness of strokes. The methods comprises: the bilateral model is improved, so that text strokes with the width within the preset range is extracted; the improved bilateral model is used for extracting stroke images from an original text image; the extracted stroke images are converted into binary images; and the binary images are treated by de-noising processing. The invention can extract all the strokes with the stroke width within the preset range, and during the image post-processing stage, a new method combining of two-level overall threshold and edge detection is provided, which can be effective in eliminating background noise while maintaining the integrity of strokes.

The invention provides a traffic scene character recognition system and recognition method on the basis of vehicle videos. According to the system, video single-frame images are pre-processed first, the most stable extreme value regions in the images are searched for and used as candidate areas of characters, and then prior knowledge is used for screening the candidate areas to obtain candidate areas conforming to the prior knowledge. The candidate areas conforming to the prior knowledge are subjected to stroke width change, a stroke width mean variance threshold value is set, some candidate areas which do not conform to stroke width features are filtered out, and finally final character areas are obtained by using a binary classifier. After the character areas are obtained, the whole character areas are divided into single characters by using a projection method and a connected domain method, and finally the single characters are sent into a well trained CNN character classifier for character recognition. In order to improve the recognition efficiency, an interesting area of the next-frame image is matched with the character area of the last-frame image by using a gray histogram for tracking and detection. The system has high positioning presicion and high recognition precision for traffic scene characters and good robustness for different lighting environments.

Apparatus and methods for dynamically rendering digital ink strokes of arbitrary stroke width while a user is drawing the ink strokes. An ink rendering system may receive sampled pen tip positions associated with a stylus or pen on a digitizer. The ink rendering system may further determine pen tip instances having particular shapes, sizes, and / or rotations, and may determine connecting shapes, preferably quadrangles, that connect between the pen tip instances. Smoothing techniques may also be used for existing digital ink.

A system and method of text detection in an image are described. A component detection module applies a filter having a stroke width constraint and a stroke color constraint to an image to identify text stroke pixels in the image and to generate both a first map based on the stroke width constraint and a second map based on the stroke color constraint. A component filtering module has a first classifier and second classifier. The first classifier is applied to both the first map and the second map to generate a third map identifying a component of a text in the image. The second classifier is applied to the third map to generate a fourth map identifying a text line of the text in the image. A text region locator module thresholds the fourth map to identify text regions in the image.

The invention discloses a text detecting method based on extraction of linear structures. A text connection region can regard as a combination of the linear structures which are in different shapes. The natural scene text detecting method comprises the following steps of roughly positioning text regions by extracting the linear structures; performing color clustering and region growth based on color distances to obtain a complete text connection region from the initial text regions; and further analyzing geometric and spatial position features of the connection region to obtain a candidate text character string. Extraction of the linear structures is based on a multi-scale Hessian matrix filtering method, a channel R, a channel G and a channel B of an original colorful image are operated respectively, and influences of illumination on the image are overcome to a certain degree by the text detecting method; the text detecting method can be used for well detecting texts with different sizes and different stroke widths by color clustering of Gaussian mixed models, region growth of color distances and a two-step color analysis method; and the robustness of the method is further improved.

The invention discloses a document image binarization method based on a support vector machine, comprising the following eight steps: color image graying, dividing a document image into blocks, improving the local contrast of image blocks, feature parameter extraction, SVM threshold classification, image block splicing, stroke width estimation, and local binarization. A color image is grayed using a minimum mean method, and the obtained gray image is of color independence. The defined local contrast not only can compensate for the influence generated by the change in image brightness, but also takes into consideration the normalized contribution of all the pixels in the neighborhood of the image to the local contrast of the image. Through use of an SVM threshold classification method, the accuracy and reliability are high. Stroke width estimation using a line-by-line scanning method is of high robustness to the resolution change of the document image. Through the method, character stroke details can be better retained, and phenomena such as ink infiltration, page stain, texture background and uneven illumination can be well suppressed while effective character foreground segmentation is achieved.

A vertical and horizontal line detection method for document images includes generating multiple binary images from the input grayscale document image based on multiple binarization thresholds, detecting horizontal and vertical lines in each of the multiple binary images independently, and merging the detection results from the multiple binary images. The line detection process for each binary image include applying an opening operation using a vertical or horizontal line as the structuring element, and removing connected components that are not vertical or horizontal lines based on a stroke width analysis. The boundaries of the detected lines are obtained using horizontal and vertical projections.

The invention discloses hardware and software algorithm designs of a columnar diode surface defect detection device. The hardware design comprises the steps of industrial camera selection, lens selection and optical platform establishment. The defect detection algorithm design of the software comprises the steps of tube body segmentation, tube body pretreatment, defective ROI segmentation, feature extraction, and decision tree classifier design. Aiming at the optical platform design, a reasonable illumination manner and a reasonable light source placement manner are tested through the optical principle and the structural properties of the object. Aiming at the design of a defect detection operator, the difficulty lies in the defective ROI segmentation and the texture feature extraction, and improved stroke width conversion and patterned gradient histogram feature extraction methods are proposed respectively; finally, the defects are classified through a decision tree classifier, the defect decision tree classifier recognition rate is close to 100%, the classification success rate reaches 96.2% and good recognition and classification results are achieved.

The invention relates to an intelligent mobile terminal scene character processing method. The method comprises the following steps: step 1, performing text coarse detection based on edges; step 2, obtaining a stroke width graph T of an input scene image I, analyzing the stroke widths and geometric characteristics of each candidate text area in a candidate text area set S, rejecting a non-text area not according with requirements, and finally outputting a positioning result graph L1; step 3, performing identification pretreatment; step 4, performing extraction operation of standard and directional element characteristics on segmented single characters; and step 5, performing fine classification based on Gabor characteristics. Compared to the prior art, the method provided by the invention has the following advantages: the accuracy is greatly improved, the recall rate is quite high, the time performance is substantially improved, and the accuracy of character identification is greatly enhanced.

The invention relates to a method for detecting text in a natural scene picture based on local width change of strokes. The method comprises the steps that 1) the natural scene picture to be detected is processed in advanced to obtain a gray scale picture, and an edge graph of the gray scale picture is obtained; 2) a corresponding binary image is generated by binarization according to the edge graph and the gray scale picture; 3) adjacent pixels of the same value are gathered to form a candidate part according to the binary image; 4) the stroke width of the candidate part is calculated to obtain the width of each pixel in stroke; 5) a filter is used to filter the candidate part according to the calculated stroke widths, and a text part is obtained by screening; and 6) the screened text parts are connected to form a text row, and a final result is displayed. The text detection method of the invention is low in calculation cost, high in efficiency and visual, and can effectively improve the text detection accuracy of the natural scene picture.

The invention provides a free scene text detection method based on affine transformation and relates to the field of image processing. According to the free scene text detection method provided by theinvention, MSER (Maximally Stable Extremal Regions) and SWT (Stroke Width Transform) are combined to realize text detection; then a vertex of a peripheral quadrilateral of each character is utilizedand is combined with an inertia principal axis to find out a quadrilateral for reflecting text distortion characteristics, so as to find out an affine parameter to carry out affine transformation; finally, detection of an image text and elimination of distortion are realized. According to the free scene text detection method provided by the invention, the accuracy of the text is remarkably improved; compared with a single text detection method based on a communication region, the recalling rate and the detection efficiency can be improved and automatic affine transformation is carried out; finally, the detection and the elimination of the image text are realized; compared with other manual affine transformation, the working efficiency is greatly improved and OCR (Optical Character Recognition) in the future is more accurate.

The invention relates to a method for locating and recognizing letters in a network animation, which comprises the following steps: fully using a structural analysis technology for the network animation to directly extract embedded coding letters in the network animation, and decomposing the network animation into basic shapes which are main elements of the network animation; then, aiming at eachbasic shape, using a two-stage letter locating method based on inspiration, combination and feature clustering of a connected domain to locate the non-embedded letter in the basic shape, and using a noise filtration method based on a stroke width column diagram to reasonably filter non-letter connected domains, thereby locating all the non-embedded letters; using the features of a gradient columndiagram to recognize the non-embedded letters; finally, returning the recognized non-embedded letters and the embedded letters to users.

The invention belongs to the technical field of Internet teaching and provides an original handwriting writing same-screen display method and system. The method comprises the steps as follows: an input client receives a handwriting input instruction of a first user; a display client receives a handwriting display instruction and stops receiving the handwriting input instruction of a second user; the input client collects position information, time information and a pressure value corresponding to the input handwriting, and obtains a stroke structure of the input handwriting and the motion speed of the input handwriting; a stroke width of the input handwriting is obtained according to the pressure value and the motion speed of the input handwriting and a data package with a predetermined format is transmitted to the display client by a server; and the display client unpacks the data package with the predetermined format and displays the stroke structure of the input handwriting and the stroke width of the input handwriting. According to the original handwriting writing same-screen display method and system, the written handwriting can be collected in real time and synchronously displayed on a plurality of display screens, so that the cooperation efficiency is improved and the paper resource is saved.

The invention, which belongs to the technical field of computer vision and particularly relates to text localization study in a complex scene, discloses a railway scene text localization method based on combination of a maximum stable extreme value region and a stroke width. On the basis of an improved histogram equalization algorithm, an original image is preprocessed, so that the contrast of an image can be improved effectively; and with an MSER algorithm, a weak target area in a railway scene is detected and a non-text area is removed based on a stroke width feature of a character, so that the false drop rate is reduced and thus problems of difficult text detection in the railway scene and difficult realization of accurate text localization can be solved. The method has the following advantages: a block sliding window search strategy is employed based on the spatial structure characteristic of the text line, so that the computational complexity can be reduced. The method can be applied to a complex railway character localization scene.

The invention discloses a handwritten text and graphic recognition system based on a paper medium. The handwritten text and graphic recognition system comprises a digital handwriting pen and a nano gridded touchpad, wherein the digital handwriting pen consists of a pen tip, a pressure sensor, a distance-measuring sensor, a data acquisition module, a storage, a battery, a color selection switch, a time-interval selection switch and a stroke-width selection switch; the gridded touchpad is formed by nano gridded microholes in a nano interval, and a pressure sensor is arranged under the microholes of the touchpad; the amounts of water and ink in the microholes are measured by the pressure sensor; and by the obtained storage amounts of the ink and the water added on each coordinate point, Chinese characters are recognized, and the profile of an ink painting is calculated to obtain accurate characters or the ink painting. The handwritten text and graphic recognition system disclosed by the invention has the advantages that direct connection with a computer is not needed, and an electronic document with Chinese characters or pictures can be formed on plain paper by handwriting.

Methods for removing vertical and horizontal lines from a document image. The horizontal line removal method includes: for a column of black pixels at each horizontal position along the line, removing them if their maximum stroke width is less than the median value of maximum stroke widths in a small window centered at that horizontal position; Remove connected components remaining in the horizontal line bounding box that do not extend significantly above or below the bounding box boundaries; and Perform closing operation to join broken pieces of character strokes caused by underline removal. This method preserves character strokes while removing underlines. The vertical line removal method includes: for vertical lines that have large height to width ratio, remove parts of such lines that are not at intersection of with horizontal or near-horizontal lines; remove all remaining connected components that touch neither left nor right boundary of the bounding box.

The invention relates to a stroke width figure based method for extracting Chinese character data from an image. The method comprises the steps of reading a colored image; clustering colors by the mean clustering algorithm to obtain a first binary image sequence; performing the edge detection algorithm and the morphological connected domain analysis method to obtain a second binary image sequence; filtering the combined sequence for the first time through a geometric filter to obtain a third binary image sequence; calculating a stroke width figure of the third binary image sequence; filtering the third binary image sequence for the second time according to the stroke width figure to obtain a fourth binary image sequence; overlapping all images in the fourth binary image sequence to obtain an extracted character result. Compared with the prior art, the method has the advantages that the distance of the color clustering algorithm can be adaptively selected by determining the image brightness value; the non-uniform illumination and other degradation phenomenon can be treated well; the traditional stroke width calculation method is improved to improve the performance of the character extracting technology.

The present invention discloses a Chinese character font vectorization method. The method comprises the following steps that: step 1, a Chinese character graph is inputted; step 2, a preprocessing stroke region is selected from the inputted Chinese character; step 3, a region selection recursive method is used to acquire all boundary point information of strokes from the selected preprocessing stroke region; step 4, the stroke type of each stroke is determined according to the boundary point information of the stroke; and step 5, as for each stroke, trajectory points on a pen moving trajectory and stroke width information perpendicular to the trajectory points on the pen moving trajectory are extracted from the boundary points, so that font vectorization is completed. The Chinese character font vectorization method is easy to implement and can efficiently and accurately vectorize the Chinese character graph.

The embodiment of the invention provides a license plate detection method and system. The method includes the steps of obtaining a license plate image, carrying out stroke width conversion is carried out on the license plate image to obtain the stroke width of each pixel point on an edge, obtaining a character area according to the stroke width of each pixel point on the edge, judging whether the character area is a license plate character area, deleting a non-license plate character area in the character area, and outputting the character area where the non-license plate character is deleted to obtain a license plate area. When an image acquisition angle of image acquisition equipment is larger than 30 degrees, the license plate area can still be detected, and the detection rate of license plates is increased.

The invention provides a stroke width extraction method and device and a character recognition method and system and solves the problem of accuracy in stroke extraction. The stroke width extraction method includes: extracting an original connected component of a stroke and allowing the component to correspond to a candidate stroke width; subjecting to the original connected component to corrosion calculation, calculating outline perimeter of the connected component before each time of corrosion, and forming a perimeter histogram; using the candidate stroke width that the connected component corresponds to after each time of corrosion calculation to perform difference calculation on the perimeter histogram so as to obtain stroke length of each candidate stroke width; forming a stroke width histogram by the stroke lengths that the candidate stroke widths correspond to; determining whether the original connected component is a character region or not according to a maximum of the stroke width histogram; if yes, determining the stroke width of the character region according to the maximum. The stroke width extraction method and device and the character recognition method and system have the advantages that stroke width extraction can be more accurate and calculation speed is increased.

The present invention relates to a night vehicle license plate positioning method based on MESR and SWT combination, and provides a novel license plate positioning method based on the MESR and SWT combination aiming at the problem that a conventional license plate positioning method can not position the license plates effectively on a night poor illumination condition or due to the vehicle speed influence. The method comprises the steps of after the contrast ratio of an original image is enhanced, carrying out the Canny edge detection and the MSER extraction, segmenting the MSER by the edge expansion and screening the segmented MSER according to the license plate character geometrical characteristics; then carrying out the SWT based on the morphological processing in the screened region; finally, aggregating a candidate region, and combining the license plate geometrical characteristics to finish the fine positioning of the license plates. By the test verification, the method is high in positioning accuracy, can position the license plates accurately on different scenes and different illumination conditions in the daytime, at the same time, can position the night low-resolution license plates effectively, and has much higher positioning accuracy and robustness than other conventional license plate positioning methods on a low-resolution condition.

The invention discloses an extracting method of strokes of a separate character in a calligraphy work. Skeleton strokes of the separate character in the calligraphy work are acquired through a skeleton crawler unit, the stroke width of each stroke of an original separate character in the calligraphy work is calculated, the minimum surrounded area of a skeleton stroke outline is acquired, relations of different types of skeleton cross points and the minimum surrounded area of the stroke outline are analyzed, shared areas are positioned, types of the shared areas are determined, and the strokes are extracted according to the types of the shared areas and the minimum surrounded area of the skeleton stroke outline. With the adoption of the method, a foundation can be laid for calligraphy character recognition, calligraphy style comparison and calligraphy work authentication based on strokes. With the adoption of the method, the strokes of the separate character in the calligraphy work can be completely extracted, and a stroke extracting method is provided for related research on and applications of stroke-based calligraphy character classification, recognition, reconstruction, animated simulation and the like.

One embodiment of the present invention sets forth a technique for stroking rendered paths. Path rendering may be accelerated when a graphics processing unit or other processor is configured to identify pixels that are within half of the stroke width of any point along a path to be stroked. The path is represented by quadratic Bèzier segments and a cubic equation is evaluated to determine whether or not each point in a conservative hull that bounds the quadratic Bèzier segment is within the stroke width.

The invention provides a multi-classifier integration-based image character recognition method. The method comprises the following steps of: converting a colored to-be-identified image into a grayscale image; carrying out binary processing on the grayscale image and segmenting an image region with character information; segmenting each Chinese character from a whole character image; extracting grid features and direction features of each Chinese character; selecting stroke density total length features to carry out first-layer rough classification by adoption of a minimum distance classifier; and respectively selecting peripheral features, the grid features and the direction features to complete second-layer classification matching by adoption of a nearest-neighbor classifier. The method has the advantages that the character recognition has relatively strong anti-jamming capability and relatively strong character local structure description capability, and is less influenced by stroke widths; by adoption of a classifier integration technology of complementing and combining the minimum distance classifier and the nearest-neighbor classifier, a system is more reliable; and the characters can be intelligently recognized, so that the adaptability of the system is improved and the recognition rate is high.