Show me your true face

Show me your true face

Problem being addressed

Even though recently deep learning based facial expression recognition methods have attracted considerable attention due to its practical importance in a variety of applications​,​ the publicly available facial expression databases typically contain a small amount of labelled data.

Solution

A novel joint deep learning of facial expression synthesis and recognition method for effective facial expression recognition. In the proposed method, which involves a two-stage learning procedure, the tasks of facial expression synthesis and recognition are simultaneously performed in a unified framework to boost the performance of each task.

Advantages of this solution

The proposed method outperforms all the other competing methods by about 2% to 15% improvement in terms of recognition rate, which further validates the effectiveness of the proposed method on large-scale databases.

Possible New Application of the Work

/static/media/automotive-1.2524d1c6.png

Automotive Industry

Smart cars that use facial expression recognition can detect facial emotions and alert the driver if he/she looks sleepy or drowsy and send an alert in advance (suggesting to take a break etc.)

/static/media/education.b165bae1.png

Education Sector

In the case of e-learning, facial expression recognition allows to study the emotions and adjust the learning technique and presentation according to the style of learner.

/static/media/others.cca5db97.png

Other Industry

Security: Accurate facial emotion recognition systems add another security layer to security at malls, airports, sports arenas, and other public venues to detect malicious intent.

Author of original research described in this blitzcard: Yan Yan, Ying Huang, Si Chen, Chunhua Shen, Hanzi Wang

question-mark-icon

Name of the author who conducted the original research that this blitzcard is based on.

Source URL: #############check-icon


search-iconBrowse all blitzcards