Improving LNMF Performance of Facial Expression Recognition via Significant Parts Extraction using Shapley Value

The non-negative Matrix Factorization (NMF) algorithms have been utilized in a wide range of real applications. NMF has been used by several researchers for its part-based representation property, especially in the facial expression recognition problems. It decomposes a face image into its essential parts (e.g. nose, lips). However, in all the previous attempts, it has been neglected that all features achieved by NMF are not required for recognition problems. For example, some facial parts do not have any useful information regarding the facial expression recognition. In this work, addressing the challenge of defining and calculating the contributions of each part, the Shapley value is used. It is applied for identifying the contribution of each feature in the classification problem, and then the effectless features are removed. Experiments performed on the JAFFE and MUG facial expression databases, as benchmarks of facial expression datasets, demonstrate the effectiveness of our approach