Offline Signature Verification Using Online Handwriting Registration
read more
Citations
Online Signature Verification on Mobile Devices
On-line signature recognition through the combination of real dynamic data and synthetically generated static data
A Model-Based Sequence Similarity with Application to Handwritten Word Spotting
Global Features for the Off-Line Signature Verification Problem
State-of-the-art in offline signature verification system
References
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data
Shape matching and object recognition using shape contexts
Dynamic programming algorithm optimization for spoken word recognition
Online and off-line handwriting recognition: a comprehensive survey
Related Papers (5)
Frequently Asked Questions (12)
Q2. What are the future works mentioned in the paper "Offline signature verification using online handwriting registration" ?
Future work includes the improvement of the precision of the recovery and the experiments on larger database with real images collected through the video-based system the authors recently developed [ 10, 25 ].
Q3. How can the authors trace the edges in G according to the order of their corresponding edges?
As the writing order of the edges in L is known, the authors can trace the edges in G according to the orders of their corresponding edges in L, thus recover the writing trajectory.
Q4. What is the purpose of this paper?
This paper focuses on offline signature verification, and their objective is to discriminate between genuine signatures and skilled forgeries which are written by careful imitation.
Q5. What is the simplest way to write a conditional probability?
According to the Hammersley-Clifford theorem [4], the conditional probability can be written into the following form [9]:pΛ(Y |X) = 1 Z(X)exp( ∑i,kλkfk(yi−1, yi, X, i)+ ∑i,kµkgk(yi, X, i)), (1)where Λ = {λk, µk} represents the set of parameters and Z(X) is a normalized factor.
Q6. What is the inference problem of the CRF?
By solving the inference problem of the CRF, the authors can determine the corresponding relations between the edges in X and the edges in G, that is, to determine the writing order of the edges in G.
Q7. What are the disadvantages of using vertices as states?
this approach has several disadvantages: 1) G may include double traced lines, so one observation may correspond to more than one state; 2) G may have complex topology which will increase the computational complexity; 3) it is difficult to account for between-edge relations if the authors use vertices/points as states, since a function of Eq.(1) includes at most two states.
Q8. What is the function F (y, i)?
Define the following function F (y, i): If i = 0, F (y, 0) = ∑kµkgk(y, X, 0);else if i > 0,F (y, i) = max y′ {F (y′, i− 1) +∑kλkfk(y′, y, X, i)}+ ∑kµkgk(y, X, i).(10)The authors can iteratively update F (y, i) using Eq. (10).
Q9. What are the advantages of using a CRF?
CRFs need a small number of training samples and allow flexible features, as it need not to specify a complete distribution for explaining observations.
Q10. What is the way to solve the optimization problem of DTW?
The optimal warping path is the one whichminimizes the above cost function,w∗ = argmin w {Dw(L, L′)}. (14)The optimization problem of DTW can be solved efficiently using dynamic programming [24].
Q11. What is the simplest way to calculate the function fk?
For each function fk, calculate its mean:f̄k = 1 N∑k∑ifk(yi−1, yi, X, i). (4)For each training pair (X, Y ), calculate:T (X, Y ) = 1 N∑Y∑ifk(yi−1, yi, X, i). (5)2. Repeat the following steps until convergence.
Q12. What is the way to find the optimal correspondence between two sequences?
DTW searches the optimal correspondence (named warping path) among the elements of two sequences by minimizing the accumulated distance.