Face Detection based on SSD and CamShift
Xizhi Hu, Bingyu Huang
Abstract
A face detection method combining SSD target detection algorithm and CamShift tracking algorithm was designed for fatigue driving detection. ResNet50 was used to replace the feature extraction network of the original SSD target detection algorithm to improve the accuracy of face location. CamShift and Kalman filter algorithm were used to track the face area to improve the detection speed and reduce the burden of system operation. The real vehicle test shows that the strategy combining SSD network and improved CamShift algorithm significantly improves the detection efficiency, and has a strong robustness to the effects of light change, occlusion loss and skin-like interference.
Topics & Concepts
Computer scienceRobustness (evolution)Artificial intelligenceKalman filterFeature extractionComputer visionFace detectionPattern recognition (psychology)Facial recognition systemChemistryBiochemistryGeneImage and Video StabilizationFace and Expression RecognitionVideo Surveillance and Tracking Methods