ACCELERATING HANDWRITTEN SIGNATURE RECOGNITION USING INTELLIGENT ALGORITHM BASED EMBEDDED SYSTEM

Year 2016, Volume: 34 Issue: 3, 393 - 405, 01.09.2016

Ali Rıza Yılmaz Burcu Erkmen Oğuzhan Yavuz

Abstract

In this work, intelligent algorithms designed on embedded hardware for signature recognition is presented. Feed forward Conic Section Function Neural Network (CSFNN) and Differential Evaluation Algorithm (DEA) are implemented on the Field Programmable Gate Arrays (FPGAs). Unified robust classifier CSFNN is applied on the preprocessed signatures for recognition purpose. DEA is used for training CSFNN in order to overcome local minimum problems. The implemented CSFNN on FPGA has the characteristics of flexible adaptable size providing various datasets. The CSFNN implementation on FPGA is realized using the 16-bit floating point arithmetic IEEE 754-2008 standard. The proposed on-chip CSFNN based signature recognition system described in VHDL has been implemented and evaluated on a high–end Virtex 7 -VC707 platform. The intelligent system embedded on FPGA is approximately 105 times faster than its equivalent software implementation.

Keywords

Conic section function neural network, FPGAs, signature recognition, differential evaluation algorithm

References

• [1] Argones Rua E., Maiorana E., Alba Castro JL., Campisi P. (2012) Biometric Template Protection Using Universal Background Models: An Application to Online Signature. IEEE Transactions on Information Forensics and Security; 7, 269-282.
• [2] Gruber C., Gruber T., Krinninger S., Sick B. (2010) Online Signature Verification With Support Vector Machines Based on LCSS Kernel Functions. IEEE Transactions on Information Forensics and Security; 40, 1088- 1100.
• [3] Ketabdar H., Richiardi J., Drygajba A. (2005) Global feature selection for on-line signature verification. In: International Graphonomics Society 2005 Conference; 26-29 June 2005, pp. 59–63, Solerno, Italy.
• [4] Erkmen B., Kahraman N., Vural RA., Yildirim T. (2010) Conic Section Function Neural Network Circuitry for Offline Signature Recognition. IEEE Transactions on Neural Networks; 21, 667- 672.
• [5] Impedovo D., Giuseppe P. (2008) Automatic Signature Verification: The State of the Art. IEEE Transactions on Systems, Man, and Cybernetics; 38, 609-635.
• [6] Al-Shoshan AI. (2006) Handwritten signature verification using image invariant and dynamic features. In: Proceedings of the 2006 International Conference on Computer Graphics, Imaging and Visualisation; 26-28 July 2006; pp. 173–176, Sydney, Australia.
• [7] Armand S., Blumenstein M., Muthukkumarasamy V. (2006) Offline Signature verification based on the Modified Direction Feature. In: Proceedings of the 18th International Conference on Pattern Recognition; 20-24 August 2006, pp. 509–512, Hong Kong.
• [8] Erkmen B., Kahraman N., Vural RA., Yildirim T. (2008) CSFNN Optimization of Signature Recognition Problem for a Special VLSI NN Chip. In: Proceedings of the 3rd Int. Symposium on Communications, Control and Signal Processing; 12-14 March 2008, pp. 1082-1085, St Julians, Malta.
• [9] Yılmaz AR.., Erkmen B., Yavuz O. (2013) The Performance of Differential Evolution Algorithm for Training CSFNN Using a Pattern Recognition Application. In: Proceedings of the 4th Int. Conference on Intelligent Control and Information Processing; 9-11 June 2013; pp. 820-823, Beijing China.
• [10] Merchant SG., Peterson GD. (2008) An evolvable artificial neural network platform for dynamic environments. In: Proceedings of the 51st Midwest Symposium on Circuits and Systems; 10-13 August 2008; pp. 77-80, Knoxville, Tennessee.
• [11] Elitas M., Yavuz O., Erkmen B. (2012) Field Programmable Gate Array implementation of Conic Section Function Neural Network: An Alternative to Analog CSFNN Circuitry. In: Proceedings of the IEEE 16th International Conference on Intelligent Engineering Systems; 13-15 June 2012; pp. 135-138, Lisbon, Portugal.
• [12] Shawash J., Selviah DR. (2013) Real-time non-linear parameter estimation using the Levenberg-Marquardt algorithm on Field Programmable Gate Arrays. IEEE Transactions on Industrial Electronics; 60, 170-176.
• [13] Gomperts A., Ukil A., Zurfluh F. (2011) Development and Implementation of Parameterized FPGA-Based General Purpose Neural Networks for Online Applications. IEEE Transactions on Industrial Informatics; 7, 78-89.
• [14] Savich AW., Moussa M., Areibi S. (2007) The Impact of Arithmetic Representation on Implementing MLP-BP on FPGAs: A Study. IEEE Transactions on Neural Networks; 18, 240 – 252.
• [15] Yılmaz AR., Yavuz O., Erkmen B. (2014) FPGA Implementation of Differential Evaluation Algorithm for MLP Training. In: Proceedings of the 2014 IEEE International Symposium on Innovations in Intelligent Systems and Applications; 23-25 June 2014, pp. 425 – 430, Alberobello Italy.
• [16] Sadykhov RK., Podenok LP., Samokhval VA., Uvarov AA. (2004) The system for handwritten symbol and signature recognition using FPGA computing. Lecture Notes in Computer Science; 3212, 447-454.
• [17] Krid M., Dammak A., Masmodi DS. (2006) FPGA Implementation of Programmable Pulse Mode Neural Network with on Chip Learning for signature application. In: Proceedings of the IEEE 13th International Conference on Electronics, Circuits and Systems; 10-13 December 2006, pp. 942-945, Nice France.
• [18] Senol C., Yildirim T. (2005) Signature Verification Using Conic Section Function Neural Network. Lecture Notes in Computer Science; 3733, 524 – 532.
• [19] Dorffner G. (1994) Unified frameworks for MLP and RBFNs: Introducing Conic Section Function Networks. Cybernetics and Systems: An International Journal 25, 511-554.
• [20] Erkmen B., Yildirim T. (2007) Obtaining Decision Boundaries of CSFNN Neurons using Current Mode Analog Circuitry. In: Proceedings of the 18th IEEE European Conference on Circuit Theory and Design; 27-30 August 2007; pp. 807-810, Seville Spain.
• [21] Erkmen B., Vural RA., Kahraman N., Yildirim T. (2013) A Mixed Mode Neural Network Circuitry For Object Recognition Application. Circuits, Systems, and Signal Processing; 32, 29-46.
• [22] Storn R., Price K. (1997) Differential evolution - a simple and efficient heuristic for global optimization over continuous spaces. The Journal of Global Optimization; 11, 341–359.
• [23] Haykin S. Neural Networks. (1994) A Comprehensive Foundation. New York, NY, USA: Macmillan College Publishing.
• [24] Rakvic RN., Ulis BJ, Broussard RP, Ives RW. (2009) Parallelizing Iris Recognition. IEEE Transactions on Information Forensics and Security 4, 812-823.
• [25] Moreno F., Alarcon J., Salvador R., Riesgo T. (2009) Reconfigurable Hardware Architecture of a Shape Recognition System Based on Specialized Tiny Neural Networks With Online Training. IEEE Transactions on Industrial Electronics; 56, 3253-3263.