Visual Speaker Authentication by a CNN-Based Scheme with Discriminative Segment Analysis
| Autor: | Jiahui Sun, Quanhai Zhang, Shilin Wang |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Scheme (programming language)
Normalization (statistics) 021110 strategic defence & security studies Authentication Computer science business.industry ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 0211 other engineering and technologies Normalization (image processing) Pattern recognition 02 engineering and technology Convolutional neural network Power (physics) Discriminative model 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence business computer Utterance computer.programming_language |
| Zdroj: | Communications in Computer and Information Science ISBN: 9783030368074 ICONIP (4) |
| DOI: | 10.1007/978-3-030-36808-1_18 |
| Popis: | Recent research shows that the static and dynamic features of a lip utterance contain abundant identity-related information. In this paper, a new deep convolutional neural network scheme is proposed. The entire lip utterance is first divided into a series of overlapping segments; then an adaptive scheme is designed to automatically examine the discriminative power and assign a corresponding weight of each segment in the entire utterance. The final authentication result of the entire utterance is determined by weighted voting of the results for all the segments. In addition, considering the various lighting condition in the natural environment, an illumination normalization procedure is proposed. Experimental results show that different segments of the same utterance have different discriminative power for user authentication, and focusing on the discriminative details will be more effective. The proposed method has shown superior performance compared with two state-of-the-art lip authentication approaches investigated. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|