A joint loss function for deep face recognition

Wang, Shanshan; Chen, Ying

doi:10.1007/s11045-018-0614-0

Cited by 3 publications

(2 citation statements)

References 72 publications

(81 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[32] The intra-class distance and the inter-class distance are widely used as indicators for feature selection of image recognition, biological information recognition, and gear wear signal recognition. [33,34] The compactness and separability of tissues status are better when the intra-class distance is smaller, and the inter-class distance is larger.…”

Section: Intra-class Distance and Inter-class Distancementioning

confidence: 99%

Identification of denatured and normal biological tissues based on compressed sensing and refined composite multi-scale fuzzy entropy during high intensity focused ultrasound treatment*

et al. 2021

View full text Add to dashboard Cite

In high intensity focused ultrasound (HIFU) treatment, it is crucial to accurately identify denatured and normal biological tissues. In this paper, a novel method based on compressed sensing (CS) and refined composite multi-scale fuzzy entropy (RCMFE) is proposed. First, CS is used to denoise the HIFU echo signals. Then the multi-scale fuzzy entropy (MFE) and RCMFE of the denoised HIFU echo signals are calculated. This study analyzed 90 cases of HIFU echo signals, including 45 cases in normal status and 45 cases in denatured status, and the results show that although both MFE and RCMFE can be used to identify denatured tissues, the intra-class distance of RCMFE on each scale factor is smaller than MFE, and the inter-class distance is larger than MFE. Compared with MFE, RCMFE can calculate the complexity of the signal more accurately and improve the stability, compactness, and separability. When RCMFE is selected as the characteristic parameter, the RCMFE difference between denatured and normal biological tissues is more evident than that of MFE, which helps doctors evaluate the treatment effect more accurately. When the scale factor is selected as 16, the best distinguishing effect can be obtained.

show abstract

Section: Intra-class Distance and Inter-class Distancementioning

confidence: 99%

Identification of denatured and normal biological tissues based on compressed sensing and refined composite multi-scale fuzzy entropy during high intensity focused ultrasound treatment*

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Moreover, some auxiliary loss [25,26] are employed to train models together with classification loss functions. Recently, some approaches focus on hard examples [27] or AutoML searching [28,29] to get better loss functions. Though these loss functions achieve better and better performance on FR, but they all depend on fulllabeled identity datasets, such as MS-Celeb-1M and Casia, and cannot treat face sequences as the training data to learn discriminative face features.…”

Section: Deep Face Recognitionmentioning

confidence: 99%

SeqFace: Learning discriminative features by using face sequences

Huang

Zhang

et al. 2021

IET Image Processing

View full text Add to dashboard Cite

Deep convolutional neural networks (CNNs) have greatly improved the Face Recognition (FR) performance in recent years. Almost all CNNs in FR are trained on the carefully labeled datasets containing plenty of identities. However, such high‐quality datasets are very expensive to collect, which restricts many researchers to achieve state‐of‐the‐art performance. In this paper, a framework, called SeqFace, for learning discriminative face features is proposed. Besides a traditional identity training dataset, the designed SeqFace can train CNNs by using an additional dataset which includes a large number of face sequences collected from videos. Moreover, the label smoothing regularization (LSR) and a new proposed discriminative sequence agent (DSA) loss are employed to enhance the discrimination power of deep face features via making full use of the sequence data. Only with a single ResNet model, the method achieves very competitive performance on several face recognition benchmarks, including LFW, YTF, CFP, AgeDB, and MegaFace. The code and model are publicly available at the website https://github.com/huangyangyu/SeqFace.

show abstract

Hierarchical fusion of visual and physiological signals for emotion recognition

Fang

Rong

Huang

2021

Multidim Syst Sign Process

View full text Add to dashboard Cite

Emotion recognition is an attractive and essential topic in image and signal processing. In this paper, we propose a multi-level fusion method to combine visual information and physiological signals for emotion recognition. For visual information, we propose a serial fusion of two-stage features to enhance the representation of facial expression in a video sequence. We propose to integrate the Neural Aggregation Network with Convolutional Neural Network feature map to reinforce the vital emotional frames. For physiological signals, we propose a parallel fusion scheme to widen the band of the annotation of the electroencephalogram signals. We extract the frequency feature with the Linear-Frequency Cepstral Coefficients and enhance it with the signal complexity denoted by Sample Entropy (SampEn). In the classification stage, we realize both feature level and decision level fusion of both visual and physiological information. Experimental results validate the effectiveness of the proposed multi-level multi-modal feature representation method.

show abstract

A joint loss function for deep face recognition

Cited by 3 publications

References 72 publications

Identification of denatured and normal biological tissues based on compressed sensing and refined composite multi-scale fuzzy entropy during high intensity focused ultrasound treatment*

Identification of denatured and normal biological tissues based on compressed sensing and refined composite multi-scale fuzzy entropy during high intensity focused ultrasound treatment*

SeqFace: Learning discriminative features by using face sequences

Hierarchical fusion of visual and physiological signals for emotion recognition

Contact Info

Product

Resources

About