Fuzzy probability based person recognition in smart environments

Biometric features are used to verify the people identity in the living places like smart apartments. To increase the chance of classification and recognition rate, the recognizing procedure contains various steps such as detection of silhouette from the gait profile, silhouette segmentation, reading features from the silhouette, classification of features and finally recognition of person using its probability value. Person recognition accuracy will be oscillated and declined due to blockage, radiance and posture variance problems. In the proposed work, the gait profile will be formed by capturing the gait of a targeted person in stipulated time to reach the destination. From the profile the silhouettes are detected using frame difference and segmented from the background using immediate thresholding and features are extracted from the silhouette using gray-level covariance matrix and optimized feature set is formed using PSO. These optimized features are fused, trained and classified using nearest neighbor support vectors. The fuzzy probability method is used for recognizing the person based on the probability value of the authentic and imposter scores. The relationship between the CMS, TPR, TNR and F-rate are calculated for 1 : 1 matcher from the gallery set. The performance of the classifiers are found to be perfect by plotting the DET graph and ROC curve. The proposed fuzzy probability theory is mingled with GLCMPSO and NSFV method for human recognition purpose. The performance of the proposed is proved to be acceptable for recognition with the optimal parameters (Entropy, SSIM, PSNR, CQM) calculation From the work, it is clear that, the rank probability is proportional to the match score value of the silhouette stored in the gallery.

Real functions and the extension of generalized probability measures II

We continue our study of the extensions of generalized probability measures. First, we describe some extensions of generalized random events (represented by classes of functions with values in [0,1]) to which generalized probability measures can be extended. Second, we study products of domains of probability and describe states on such products. Third, we show that the events in IF-probability, introduced by B. Riečan, form a suitable category isomorphic to a subcategory of the category of fuzzy random events. Consequently, IF-probability can be interpreted within fuzzy probability theory. We put forward some problems related to the extensions of probability domains and hint some applications.

On D-posets of fuzzy sets

D-posets of fuzzy sets constitute a natural simple mathematical structure in which relevant notions of generalized probability theory can be formalized. We present a classification of D-posets leading to a hierarchy of distinguished subcategories of D-posets related to probability and study their relationships. This contributes to a better understanding of the transition from classical probability theory to fuzzy probability theory. In particular, we describe the transition from the Boolean cogenerator {0, 1} to the fuzzy cogenerator [0, 1] and prove that the generated Łukasiewicz tribes form an epireflective subcategory of the bold algebras.

Kinematic accuracy reliability research of a novel exoskeleton with series-parallel topology

A novel type of lower extremity exoskeleton with series-parallel topology was described. Position analysis of a simplified leg was conducted to provide a basis of the kinematic accuracy reliability. Considering both assembly-machining errors and driving errors, the fuzzy reliability of kinematic accuracy was obtained based on the fuzzy probability theory. With a set of optimized length parameters and joint ranges of motion given, the fuzzy reliability of space distribution and the average fuzzy reliability were respectively calculated and shown in figures. These analyses of kinematic accuracy reliability provide a unique method to judge, analyze, and design exoskeleton mechanisms.