Eyelid key point accurate positioning method based on deep convolutional neural network

• Main Authors: TENG TONG, MAO YUNFENG, CHAO JINGJING, SHEN WENZHONG, SONG TIANSHU
• Format: Patent
• Language: Chinese; English
• Published: 25.09.2018
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention relates to an eyelid key point accurate positioning method based on a deep convolutional neural network. The method uses the pre-trained deep convolutional neural network to determine boundary points between an eyelid and an iris, and then fits the boundary points to acquire an eyelid boundary line. The method used for acquiring the deep convolutional neural network comprise the steps that (1) image preprocessing is carried out to acquire an iris region as an image of a region of interest; (2) the boundary points between the eyelid and the iris are manually marked as marker points on the image of the region of interest; (3) an image training set is acquired based on the image of the region of interest; and (4) the deep convolutional neural network is established, and the image training set is used as input and sent into the deep convolutional neural network to complete the training. Compared with the prior art, the method provided by the invention has the advantages of fast positioning speed, hi