Abstract
We propose a novel and general framework called the multithreading cascade of Speeded Up Robust Features (McSURF), which is capable of processing multiple classifications simultaneously and accurately. The proposed framework adopts SURF features, but the framework is a multi-class and simultaneous cascade, i.e., a multithreading cascade. McSURF is implemented by configuring an area under the receiver operating characteristic (ROC) curve (AUC) of the weak SURF classifier for each data category into a real-value lookup list. These non-interfering lists are built into thread channels to train the boosting cascade for each data category. This boosting cascade-based approach can be trained to fit complex distributions and can simultaneously and robustly process multi-class events. The proposed method takes facial expression recognition as a test case and validates its use on three popular and representative public databases: the Extended Cohn-Kanade, MMI Facial Expression Database, and Annotated Facial Landmarks in the Wild database. Overall results show that this framework outperforms other state-of-the-art methods.
Highlights
And simultaneously learning highly discriminative multiclass classifiers with local image features is one of the most significant challenges to computer vision researchers, because they are critical infrastructures for recognition engines; these researches appear of great importance
The proposed learning model is applied to facial expression recognition (FER), and while it is derived from AdaBoost [1], it is a novel, multi-class, simultaneous cascade, i.e., a multithreading cascade
3 The proposed method Our proposed framework has these components: SURF features for local patch description; logistic regressionbased weak classifiers, which are combined with the area under the receiver operating characteristic (ROC) curve (AUC) [18] as a single criterion for cascade convergence testing; and a multithreading cascade for boosting training that can process multiple categories
Summary
And simultaneously learning highly discriminative multiclass classifiers with local image features is one of the most significant challenges to computer vision researchers, because they are critical infrastructures for recognition engines; these researches appear of great importance. 2 Related work Recently, mainstream FER approaches are based on effective local descriptors or facial action units Local descriptors such as local binary pattern on three orthogonal planes (LBP-TOP) [12], HOE [13], and histograms of oriented gradients (HOG) 3D [14] are extracted from the local facial cuboid to obtain a representation of a certain length independent of time resolution. The effective use of local descriptors to represent complex expressions has been an ongoing problem Another approach is adopted for processing facial action areas. We propose a novel and general learning framework that contains robust classifiers as well as high-quality local feature descriptors, and the technical details are discussed
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
