Research on Inception Module Incorporated Siamese Convolutional Neural Networks to Realize Face Recognition

Abstract

Face recognition is an active research subject of biometrics due to its significant research and application prospects. The performance of face recognition can be affected by a series of uncontrollable factors, such as illumination, expression, posture and occlusion, which restricts its real-world applications. Therefore, improving the robustness of face recognition to environmental changes became an urgent problem. In this paper, a simplified deep convolutional neural network structure having high robustness under unlimited conditions is designed for face recognition. This structure can improve training speed and face recognition accuracy, and be suitable for small-scale data sets. Inception Module Incorporated Siamese Convolutional Neural Networks (IMISCNN) is developed based on effective reduction of external interference and better features extraction by adopting the Siamese network structure. A cyclical learning rate strategy is also introduced in IMISCNN for better model convergence. Compared to classical face recognition algorithms, such as PCA, PCA and SVM, CNN, PCANet, and the original SNN et al. The accuracy of IMISCNN in CASIA-webface and Extended Yale B standard face database is 99.36% and 99.21%, respectively. Its feasibility and effectiveness have been verified in our experiments. INDEX TERMS Cyclical learning rate, face recognition, inception module, Siamese convolutional neural networks.

A Survey on Symmetrical Neural Network Architectures and Applications

“…As one can see from the figure below, the most popular methods were mainly based on the work of [7] (references are indicated by arrows). [29]; Heidari and Fouladi-Ghaleh (2020) [30]. The summaries of these works are provided in Section 4.3.…”

“…The similarity metric based convolutional neural network (SMCNN) solving kinship verification tasks was first proposed in [26]. This network takes as input two images of close relatives' faces (for example, father-son, father-daughter, mother-son, mother- [29]; Heidari and Fouladi-Ghaleh (2020) [30]. The summaries of these works are provided in Section 4.3.…”

“…In [29] it was proposed to use the Inception module incorporated Siamese convolutional neural network (IMISCNN) and the cyclical learning rate policy to train it. The IMISCNN architecture comprises two CNNs with a single Inception module containing more information due to different receptive fields.…”

Review of Wide-Baseline Stereo Image Matching Based on Deep Learning

“…Figure 8a shows the traditional "detect-then-describe" model, that is, SuperPoint [81], which is a representative model of this type, and Figure 8b shows the D2Net "describe-and-detect" model. In contrast to a Siamese or multi-branch network structure [87], D2Net adopts a single-branch network architecture, and the feature location and descriptor information of the image are stored in high-dimensional feature channels, which is thus more conducive to obtaining stable and efficient matches. However, D2Net must extract dense descriptors in the process of using high-level semantic information, which reduces the accuracy and efficiency of feature detection.…”

MFCosface: A Masked-Face Recognition Algorithm Based on Large Margin Cosine Loss

“…They calculated the average facial features of each identity and performed identity recognition. Xu et al [21] proposed a Siamese convolutional neural network for facial recognition based on the Siamese convolutional neural network and the Inception module. In the absence of masked-face datasets, most methods only simulate mask occlusion by adding random noise or black pixels, which makes their abilities with real mask occlusion questionable.…”