Consensual Collaborative Training And Knowledge Distillation Based Facial Expression Recognition Under Noisy Annotations

Gera, Darshan; Balasubramanian, Siva

doi:10.14445/22315381/ijett-v69i7p231

Cited by 6 publications

(2 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the pooling layer input y, the pooling process is [9] pool � down max y i,j 􏼐 􏼑 􏼐 􏼑, i, j ∈ p, (5) where y represents the pooling layer input element, and do wn is the downsampling process, which retains the maximum value in the pooled area.…”

Section: Training Deep Convolutional Neural Networkmentioning

confidence: 99%

“…About the expression recognition since the beginning of the 21th century, based on a simple knowledge distillation scheme, Gera D uses a single network to complete label inference on large-scale facial expression datasets, and the proposed framework demonstrates its effectiveness and comprehensiveness in the noisy FER dataset [5]. Li et al fused energy feature vectors and facial feature vectors in their research study, used support vector machine (SVM) to classify the fused feature vector, and finally proved that the proposed method has higher accuracy and stronger generalization transformation ability [6].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Construction of Facial Expression Recognition Technology of Intelligent Service Robot Based on Nanofiber Flexible Wearable Electronic Skins

Liu

2022

Advances in Materials Science and Engineering

View full text Add to dashboard Cite

Recognizing human features is a highly significant and popular research topic in academia today, especially with regard to human intervention and computer intelligence. If you want your computer to have real intelligence, then the computer must first learn to understand human feelings, and facial expressions often contain rich emotional information. Clearly, recognizing human features is indispensable for the future intelligent development of computers. There are always a few components and many researchers are involved. Compared with the traditional facial expression recognition method, which needs to manually define features and artificially perform appropriate feature extraction, a great advantage of deep learning is to remove the disadvantages that require a lot of human intervention, thereby reducing the uncertainty of expression recognition. In recent years, deep processing has led to outstanding gains in numerous domains, including robotics, computer graphics, and natural language processing. Therefore, this article applies a popular recognition algorithm based on deep learning, combined with a nanofiber flexible wearable electronic skin, and applies it to the field of facial expression recognition. Functional expression identification is an essential goal of the external analysis of human features. Mature facial expression recognition technology has great application values.

show abstract

Section: Training Deep Convolutional Neural Networkmentioning

confidence: 99%

mentioning

confidence: 99%