Person Re-identification Based on Deep Learning Algorithms with Manual Extracted Colour and Texture Features
Xiang‐Qian Chang, Zhihao Su, Ni Li, Chenyang Liu, Huimin Xue, Yunxi Nie
Abstract
Person re-identification technology is an important research direction in the field of computer vision and intelligent transportation. Deep learning algorithms have better performance than traditional feature extraction algorithms in recent years. The paper proposes a hybrid deep learning algorithm with manual colour and texture features. The proposed hybrid model in the paper has 3-5 % higher accuracy than the single deep learning algorithms in the two mainstream databases. The experiment proves that the combination between deep learning algorithms and traditional algorithms has great significance.
Topics & Concepts
Computer scienceArtificial intelligenceDeep learningAlgorithmIdentification (biology)Feature extractionMachine learningField (mathematics)Statistical classificationPattern recognition (psychology)MathematicsBotanyBiologyPure mathematicsVideo Surveillance and Tracking MethodsAutomated Road and Building ExtractionIoT and GPS-based Vehicle Safety Systems