<p>To solve the problem of poor face recognition performance in unrestricted environments. A face recognition algorithm based on improved residual IR-ResNet-SE is designed. Firstly, the IR structure is added to the 34-layer residual network to reduce the variability of different features; Secondly, we add the channel attention module to increase the weight of important channel features; Finally, the Arcface loss function is used to improve the classification ability of the model. The LFW, AgeDB, and AR datasets reflect unrestricted factors such as pose, age, expression, occlusion, and illumination. The algorithm proposed in this paper is experimented on these three datasets. The experimental results show that the IR-ResNet-SE algorithm proposed in this paper can achieve 99.74% accuracy in the dataset LFW. And it has excellent robustness in face recognition under unrestricted conditions.</p> <p>&nbsp;</p>

<p>Facial expression recognition is widely used, but there are some problems such as complex scenes, lack of data sets and low recognition rate. In this paper, we construct a new network model and name it RNFC. The RNFC network adopts 6 improved residual blocks to extract features. Features are passed into the fully connected layer by flattening the data, and Dropout techniques are introduced between the fully connected layers to prevent overfitting of the model. Based on the pytorch framework, we use a cross-entropy loss function to improve the training speed of the network. And perform denoising and enhancement pre-processing on the FER2013 dataset. The RNFC network is trained and tested on the pretreated FER2013. It has a higher recognition rate than classical networks such as VGGnet19 and ResNet18.</p> <p>&nbsp;</p>