ONLINE SIGNATURE RECOGNITION AND VERIFICATION USING (SURF) ALGORITHM WITH SVM KERNELS | ||||
| Journal of Al-Azhar University Engineering Sector | ||||
| Article 24, Volume 13, Issue 49, October 2018, Page 1332-1344 PDF (701.82 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/auej.2018.18939 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
| Ali Khaleel Ibrahim Hamadly; Hossam Eldin Hassan Abdel Munim; Hoda Korashy Mohamed | ||||
| Computer and Systems Engineering, Department Ain Shams University | ||||
| Abstract | ||||
| One important way to identify the signer is the personal signature. The operation of recognize the signature and identify it, is very important. This process may be done either offline or online. This paper explains the online technique. Features extraction, pattern matching and Images processing are techniques used for signature confirmation. Speed up robust features (SURF) is an algorithm that uses the image of local feature with ability for matching images. SURF recognizes, describes and extracts the local feature of forged signature from the image. SURF algorithm provides fast and accurate comparison tool that can work under different lights, visions and rotation situations to check if the person signature is original or forged. The features extracted from the SURF algorithm are entered into Bag-of-word features. The features of bag-of-word are used inside multiclass Support Vector Machine (SVM) classifies. SURF with SVM kernels gives an accuracy of 98.75%. | ||||
| Keywords | ||||
| SURF; SVM; Feature Extraction | ||||
|
Statistics Article View: 294 PDF Download: 761 |
||||
