A new method that amends the label distribution of each facial image byleveraging correlations among expressions in the semantic space . By utilizing label distribution learning, a probability distribution is assigned for a facial image to express a compound emotion, which effectivelyimproves the problem of label uncertainties and noises occurred in one-hotlabels . In practice, correlations among emotions areinherently different, such as surprised and happy emotions are more possiblysynchronized than surprised and neutral . Experimental results demonstratethe proposed method is more effective than compared state-of-the-art methods . It indicates the correlation may becrucial for obtaining a reliable label distribution. It indicates that the correlation is crucial for obtaining the correlation . It is more likely to be used to accurately predict the labels of an image that expresses an emotion that is associated with a certain type of a certain amount of a particular emotion .

Author(s) : Shasha Mao, Guanghui Shi, Licheng Jiao, Shuiping Gou, Yangyang Li, Lin Xiong, Boxin Shi

Links : PDF - Abstract

Code :
Coursera

Keywords : label - distribution - emotion - correlation - facial -

Leave a Reply

Your email address will not be published. Required fields are marked *