Bimodal person identification based on speech and hand signature recognition /
Automatic person identification systems today become essential and highly demanded. During ancient time, persons' characteristics such as face and voice were widely used to recognize each other and identify persons in criminal situations. However, these manual solutions are limited and highly d...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Engineering, International Islamic University Malaysia,
2012
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Automatic person identification systems today become essential and highly demanded. During ancient time, persons' characteristics such as face and voice were widely used to recognize each other and identify persons in criminal situations. However, these manual solutions are limited and highly dependent upon capabilities of human. These manual solutions are transformed today into automatic solutions that are capable enough to perform person identification tasks effectively. These automatic solutions believe that each person has his/her own distinct characteristics, which make· biometrics as effective person's identifiers and excellent substitutions to traditional person identification methods such as identification card, password, and Personal Identification Number (PIN).This research work aims to develop automatic person identification systems based on three approaches including speech, hand signature, and bimodal biometric technologies. In order to develop the speech system, the Mel-Frequency Cepstral Coefficient (MFCC) and the Vector Quantization (VQ) are used, whereas the hand signature system is based on global features and Multilayer Perceptron (MLP) architecture for. Artificial Neural Network (ANN). The bimodal system is based on fusing the speech and hand signature systems at decision level. In order to develop the systems, speech and hand signature data are collected from 100 persons (50 male and 50 female). The speech database includes recordings for the persons' preferred usemames, which are recorded 30 times by each person and resulted in a total of 3000 recordings. In addition, each person signed 30 times, which resulted in a total of 3000 hand signature images. Both databases are used in developing the bimodal system. Overall experimental results show that the bimodal system is the best solution and better than the single modality systems, whereby the bimodal system obtained an average recognition rate of 96.40%, whereas the speech system and the hand signature system obtained average recognition rates of 92.60% and 75.20%, respectively. |
|---|---|
| Item Description: | Abstracts in English and Arabic. " A dissertation submitted in fulfilment of the requirement for the degree of Master of Science (Communication Engineering)."--On t.p. |
| Physical Description: | xvii, 143 leaves : illustrations ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 120-127). |