Representation and recognition of handwritten

digits using deformable templates

A presentation of a case study on title

Definition

• Optical Character Recognition (OCR) is the mechanical or electronic conversion of images of typewritten or printed text into machine-encoded text.

Introduction to OCR

Introduction

• Automatic Recognition of hand-printed characters has long been a goal of many research efforts in the pattern recognition field.

• This working prototype system can detect handwritten digits from a scanned image of an input form by using deformable templates technique.• Many methods for extraction or extracting features from character

images have been proposed, the proposed features include counts of topological features crossing , endpoints, holes ,etc.. and various mathematical moments.

Objective

• To recognize handwritten digits in real works for autonomous machine processing.• Primary Performance measures are classification accuracy and

recognition speed.• Deformable templates: where an image deformation is used to match

an unknown image against a database of known images.• The goal of this paper is to investigate the deformation of character

image outlines as a source of information for recognition

Deformable Templates• two characters are matched by deforming the contour of one to fit the

edge strengths of the other and dissimilarities measure is derived from the amount of deformation needed, the goodness of fit of the edges the interior overlap between the deformed shapes.• Deformable templates: where an image deformation is used to match

an unknown image against a database of known images.• classification using the minimum dissimilarity results in recognition

rates up to 99.25% on 2.000 character subset of NIST “special database”.• Additional experiments on independent test data were done to

demonstrate the robustness of the method.

Deformable Templates• We present our deformation model, discuss the use of this for feature

extraction and present result with this method on a 2.000 image NIST SD-1 hand digit data.

• The NIST images , we have worked with are from the FL-3 distribution a subset of SD-1 containing approximate 3.500 digit images.• We have another dataset training for classifying 2.000 image from IBM.

Deformable Model For Recognition• Samples which cannot be satisfactorily assigned to a class in the way are

passed to a slower relaxation matching algorithm which uses deformation to match the sample to each template.• They report a 93.15% recognition rate on 2.000 sample database token

from USPS ZIP code images.• Model characters with a spline and assume that the spline parameters

have a multivariate Gaussian distribution.• Bayesian approach is the used to determine the character class with the

model parameters as prior and image data parameters as likelihood this achieved a 95.4% recognition rate on NIST SD-1 hand-printed digit set .

Deformable Model For Recognition• Model digits as ink-generating Gaussian beads string along a spline

outline characters are matched through deformation of the spline and adjustment of the bead parameters their best results report has 99.00% recognition accuracy on 2.000 character set with no rejection.• For digit recognition system based on binary template matching the

authors report recognition rates in the range of 94.03%-96.39 with error rates in the range 0.54%-1.05%.

methodology• The basic goal is to determine the dissimilarity between two digit

images using a deformable template approach.

Multidimensional scaling for feature extraction

• Is a well-known technique to obtain an appropriate representation of the patterns from the given proximity matrix, given an n*n input matrix of inter-pattern distances, multidimensional scaling creates an n*d pattern matrix, embedding the n patterns as points in d dimensional space, trying to keep the distance between patterns as close to the input dissimilarity matrix as possible.

Feature Extraction• In feature extraction stage each character is represented as a feature

vector, which becomes its identity. The major goal of feature extraction is to extract a set of features, which maximizes the recognition rate with the least amount of elements. • Due to the nature of handwriting with its high degree of variability

and imprecision obtaining these features, is a difficult task. Feature extraction methods are based on 2 types of features: • Statistical• Structural

Statistical Features

• Representation of a character image by statistical distribution of points takes care of style variations to some extent.

• The major statistical features used for character representation are: • Zoning• Projections and profiles• Crossings and distances

Zoning• The character image is divided into NxM zones. From each zone

features are extracted to form the feature vector. The goal of zoning is to obtain the local characteristics instead of global characteristics

Zoning – Density Features• The number of foreground pixels, or the normalized number of

foreground pixels, in each cell is considered a feature.

Darker squares indicate higher density of zone pixels.

Projection Histograms• The basic idea behind using projections is that character images,

which are 2-D signals, can be represented as 1-D signal. These features, although independent to noise and deformation, depend on rotation.• Projection histograms count the number of pixels in each column and row of a

character image. Projection histograms can separate characters such as “m” and “n” .

Profiles• The profile counts the number of pixels (distance) between the

bounding box of the character image and the edge of the character. The profiles describe well the external shapes of characters and allow to distinguish between a great number of letters, such as “p” and “q”.

Crossings and Distances• Crossings count the number of transitions from background to foreground pixels

along vertical and horizontal lines through the character image and Distances calculate the distances of the first image pixel detected from the upper and lower boundaries, of the image, along vertical lines and from the left and right boundaries along horizontal lines.

Structural Features• Structural features are based on topological and geometrical

properties of the character.• such as aspect ratio, cross points, loops, branch points, strokes and

their directions, inflection between two points, horizontal curves at top or bottom, etc…