Objective To determine the audiological characteristics in 832 deaf children with biallelic causative mutations in GJB2 ,SLC26A4 gene .Methods The 832 patients received deafness gene screening ,553 were GJB2 gene biallelic causative mutations ,279 were SLC26A4 gene biallelic causative mutations .Patients were divided into four groups according to ages of hearing loss onset :<1 ,1~3 ,3~6 ,6~12 years old ,and the audiological character-istics and prevalence of GJB2 ,SLC26A4 gene mutations at different ages of onset .Results The prevalence of GJB2 gene mutations at four groups was 37 .97% (210/553) ,38 .34% (212/553) ,16 .27% (90/553) ,7 .41% (41/553) ,re-spectively ;the prevalence of SLC26A4 gene mutations at four groups was 25 .45% (71/279) ,44 .80% (125/279) , 20 .07% (56/279) ,9 .67% (27/279) ,respectively .The difference between GJB2 and SLC26A4 gene was significant(P=0 .001) .The prevalence of profound hearing loss with GJB2 gene mutations at four groups were 66 .67% (140/210) ,61 .32% (130/212) ,47 .78% (43/90) ,41 .46% (17/41) ,respectively .The difference was significant (P=0 .004) ,while the difference in 279 patients with SLC26A4 gene mutations was not statistically significant (P= 0 . 083) .Conclusion The age of hearing loss onset in patients with biallelic causative mutations in GJB 2 or SLC26A4 gene refers to 0~3 years -old ,hearing loss in patients with GJB2 ,SLC26A4 gene mutations gives priority to pro-found .The age of hearing loss onset is smaller ,the ratio of profound hearing loss is higher .Patients with severe and profound hearing impairment should be performed the genetic testing when the age of onset under 12 .

We propose a novel palm-vein recognition model based on the end-to-end convolutional neural network. In this model, the convolutional layer and the pooling layer were alternately connected to extract the image features, and the categorical attribute was estimated simultaneously via the neural network classifier. The classification error was minimized via the mini-batch stochastic gradient descent algorithm with momentum to optimize the feature descriptor along with the direction of the gradient descent. Four strategies including data augmentation, batch normalization, dropout, and L2 parameter regularization were applied in the model to reduce the generalization error. The experimental results showed that for classifying 500 subjects form PolyU database and a self-established database, this model achieved identification rates of 99.90% and 98.05%, respectively, with an identification time for a single sample less than 9 ms. The proposed approach, as compared with the traditional method, could improve the accuracy of palm vein recognition in clincal applications and provides a new approach to palm vein recognition.
Algorithms ; Databases, Factual ; Hand ; blood supply ; diagnostic imaging ; Humans ; Neural Networks (Computer) ; Veins ; diagnostic imaging