We propose a novel adaptive fast learning (AFL) algorithm for two-dimensional principal component analysis (2DPCA) in this paper. As opposite to conventional PCA which is based on 1D data vectors, 2DPCA is based on 2D image matrices and thus has higher accuracy than conventional PCA when applied to applications such as face recognition, facial expression recognition, palmprint recognition, etc. Our proposed AFL algorithm simultaneously estimates both eigenvectors and corresponding eigenvalues, and then adaptively sets the learning rate parameters of neurons to ensure all neurons learning with almost the same fast speed. Requiring no image covariance matrix evaluation, the desired multiple eigenvectors of 2DPCA can thus be learned effectively in the form of weight vectors of neurons. The proposed AFL algorithm can also be applied to learning for T-2DPCA. Simulation experiments performed on face database such as the YaleB database clearly demonstrate that the proposed AFL algorithm performs very well and thus is a very effective computational tool for both 2DPCA and T-2DPCA.

Human eye state identification can be applied not only to monitoring of the drowsiness of a human car driver but also to medical treatment facilitating system for monitoring neonate or stuporous patient. Once the patient awake and open his eyes, human eye state identification system can notify nurses to take care of the patient. In this work, we propose an intelligent human eye state identification algorithm based on 2DPCA and skin color. Adaboost face detection function of OpenCV is first adopted to detect the human faces in color images acquired from camera. Then, we develop a more precise HSV skin color model and use it to eliminate the false alarms in the previous stage. Next, a heuristic segmentation method based on skin color and face geometry is proposed to segment the region of eyes, from which 2DPCA is then adopted to extract the features and identify the opening or closing state of eyes. We study three kinds of 2DPCA, i.e. 2DPCA, T-2DPCA and (2D)2PCA, and compare their performance. Experimental results reveal that our algorithm can achieve over 90% accuracy rate.

It is difficult to objectively and quantitatively judge image quality by a single criterion, such as contrast. In general, excessive contrast enhancement easily leads to a loss of image quality. Thus, it easily gives a wrong evaluation to rank image quality according to contrast values. In order to achieve a consistent result with human vision perception, balancing multi-criteria will be a feasible approach. Therefore, we propose a multi-criteria image quality evaluation scheme for ranking seven existing contrast enhancement methods. The scheme applies four criteria to a newly proposed way of computing a grey relational grade (GRGd), called the consistent grey relational grade (CGRGd). Experimental results show that our proposed CGRGd do provides a very effective mechanism to choose the best method for a specific purpose.