Method and system for determining vanishing point candidates for projective correction

A method for projective correction of an image containing at least one text portion that is distorted by perspective, the method comprising the steps of:

• image binarization, wherein said image is binarized;
• performing connected component analysis, wherein pixel blobs are detected in said at least one text portion of said binarized image and wherein for each of said pixel blobs a position determining pixel is selected on a pixel blob baseline of the pixel blob, said position determining pixel defining the position of the pixel blob in the binarized image;
• horizontal vanishing point determination, comprising estimating text baselines by means of said position determining pixels of said pixel blobs and determining at least one horizontal vanishing point candidate of said at least one text portion by means of said text baselines;
• vertical vanishing point determination, comprising estimating vertical text blob lines which each correspond to the direction of a selected one of said pixel blobs, selected by a blob filtering algorithm on the text portion of the image, and determining at least one vertical vanishing point candidate of said at least one text portion by means of said vertical text blob lines;
• wherein at least one of said horizontal and vertical vanishing point determination comprises the steps of:
• transforming each of said estimated text lines to a data point in a homogenous coordinate system;
• assigning a confidence level to each of the data points;
• grouping a number of data points having a confidence level above a predetermined threshold into a priority sample array;
• clustering the data points in the priority sample array into a number of sample groups, wherein each sample group comprises at least two data points;
• assigning a group confidence value to each sample group on the basis of at least the confidence level assigned to each data point in the sample group;
• applying a RANSAC algorithm to determine among said data points a set of inliers with respect to a first fitted line, wherein said RANSAC algorithm is initiated with the sample group having the highest group confidence value in the priority sample array;
• and estimating said at least one vanishing point candidate from the text lines corresponding to said set of inliers;
• and projective correction, wherein said perspective in said image is corrected on the basis of a horizontal vanishing point selected among said at least one horizontal vanishing point candidate and a vertical vanishing point selected among said at least one vertical vanishing point candidate.

A system for projective correction of an image containing at least one text portion that is distorted by perspective, the system comprising at least one processor and an associated storage containing a program executable by means of said at least one processor and comprising:

• first software code portions configured for image binarization, which when executed binarize said image;
• second software code portions configured for connected component analysis, which when executed detect pixel blobs in said at least one text portion of said binarized image and for each of said pixel blobs select a position determining pixel on a pixel blob baseline of the pixel blob, said position determining pixel defining the position of the pixel blob in the binarized image;
• third software code portions configured for horizontal vanishing point determination, which when executed estimate text baselines by means of said position determining pixels of said pixel blobs and determine at least one horizontal vanishing point candidate of said at least one text portion by means of said text baselines;
• fourth software code portions configured for vertical vanishing point determination, which when executed estimate vertical text blob lines which each correspond to the direction of a selected one of said pixel blobs, selected by a blob filtering algorithm on the text portion of the image, and determine at least one vertical vanishing point candidate of said at least one text portion by means of said vertical text blob lines;
• wherein at least one of said third and fourth software code portions is configured for performing the steps of:
• transforming each of said estimated text lines to a data point in a homogenous coordinate system;
• assigning a confidence level to each of the data points;
• grouping a number of data points having a confidence level above a predetermined threshold into a priority sample array;
• clustering the data points in the priority sample array into a number of sample groups, wherein each sample group comprises at least two data points;
• assigning a group confidence value to each sample group on the basis of at least the confidence level assigned to each data point in the sample group;
• applying a RANSAC algorithm to determine among said data points a set of inliers with respect to a first fitted line, wherein said RANSAC algorithm is initiated with the sample group having the highest group confidence value in the priority sample array;
• and estimating said at least one vanishing point candidate from the text lines corresponding to said set of inliers;
• and fifth software code portions configured for performing projective correction, which when executed correct said perspective in said image on the basis of a horizontal vanishing point selected among said at least one horizontal vanishing point candidate and a vertical vanishing point selected among said at least one vertical vanishing point candidate.

A non-transient storage medium on which a computer program product is stored comprising software code portions in a format executable on a computer device and configured for performing the following steps when executed on said computer device:

• image binarization, wherein said image is binarized;
• performing connected component analysis, wherein pixel blobs are detected in said at least one text portion of said binarized image and wherein for each of said pixel blobs a position determining pixel is selected on a pixel blob baseline of the pixel blob, said position determining pixel defining the position of the pixel blob in the binarized image;
• estimating a number of text lines in a Cartesian coordinate system, each text line representing an approximation of a horizontal or vertical text direction of said text portion, on the basis of the position determining pixels;
• transforming each of said text lines to a data point in a homogenous coordinate system;
• assigning a confidence level to each of the data points;
• grouping a number of data points having a confidence level above a predetermined threshold into a priority sample array;
• clustering the data points in the priority sample array into a number of sample groups, wherein each sample group comprises at least two data points;
• assigning a group confidence value to each sample group on the basis of at least the confidence level assigned to each data point in the sample group;
• applying a RANSAC algorithm to determine among said data points a set of inliers with respect to a first fitted line, wherein said RANSAC algorithm is initiated with the sample group having the highest group confidence value in the priority sample array;
• estimating at least one vanishing point candidate from the text lines corresponding to said set of inliers.