Speed Invariant Human Gait Authentication Based on CNN
| Autor: | K. R. Radhika, K. Ambika |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Authentication
Biometrics Computer science business.industry ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Pattern recognition Convolutional neural network Silhouette Identification (information) ComputingMethodologies_PATTERNRECOGNITION Gait (human) Multilayer perceptron Covariate Artificial intelligence business ComputingMethodologies_COMPUTERGRAPHICS |
| Zdroj: | Lecture Notes in Networks and Systems ISBN: 9783030847593 |
| DOI: | 10.1007/978-3-030-84760-9_68 |
| Popis: | Human Gait is used as a biometric property for person identification and it plays an important role in security and video surveillance applications. Gait is a potential solution for surveillance systems to ascertain identity at a distance from camera. Silhouette-based gait is widely used in the current gait recognition community due to their effectiveness and efficiency, but they are subject to changes in covariate conditions. The following algorithm classifies a gait irrespective of the speed variation and covariate like back pack. Convolutional Neural Network Multilayer Perceptron (MLP) based approach is used for the classification of walking gaits for different subjects for authentication in their respective classes. This research incorporates Gait Dataset-C from CASIA of Center for Biometric and Security Research (CBSR) by including still video frames from different variants of walking gait silhouettes for subjects under consideration. In the proposed method experimental results show the classification accuracy of 99.3%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|