The effect of microwave coupled hot air drying on rehydration ratio (RR) and total sugar content (TSC) of Chinese yam was investigated. Single factor test and response surface methodology were used for process parameter optimization with hot air temperature, hot air velocity, slice thickness, and microwave power density as variables and RR and TSC of dried products as responses. The effect of variables on RR followed the order: slice thickness > hot air temperature > microwave power density > hot air velocity. The effect of variables on TSC followed the order: slice thickness > microwave power density > hot air velocity > hot air temperature. The optimized process parameters were hot air velocity of 2.5 m/s, hot air temperature of 61.7 °C, slice thickness of 8.5 mm, and microwave power density of 5.9 W/g. Under the optimal conditions, the predicted values of RR and TSC were 1.90 g/g and 5.74 g/100 g, respectively, which is very close to corresponding actual values (1.83 g/g and 5.72 g/100 g). The desirability of 0.913 further validated the effectiveness of the model. The findings from this work may apply to other agricultural products.

Microwave coupled with hot air drying kinetics and characteristics of hawthorn slices at different drying hot air temperatures, hot air velocities, and microwave power densities was investigated. The research results showed that drying occurred mainly in the falling rate period and in the accelerating period. Twelve mathematical models were selected to describe and compare the drying kinetics of hawthorn slices. By comparing three criterions including correlation coefficient, chi-square, and root mean square error, we determined that Weibull distribution model obtained the best fit and could best predict the experimental values. Consequently, Weibull distribution model could be used to aid dryer design and promote the efficiency of dryer operation by simulation and optimization of the drying processes. Moisture transfer from hawthorn slice was described by applying Fick’s second law and the effective diffusivity values were calculated by simplified Fick’s second law. The variable law of effective diffusivity values was consistent with the variable law of moisture ratio.

Face recognition has advanced considerably with the availability of large-scale labeled datasets. However, how to further improve the performance with the easily accessible unlabeled dataset remains a challenge. In this paper, we propose the novel Unknown Identity Rejection (UIR) loss to utilize the unlabeled data. We categorize identities in unconstrained environment into the known set and the unknown set. The former corresponds to the identities that appear in the labeled training dataset while the latter is its complementary set. Besides training the model to accurately classify the known identities, we also force the model to reject unknown identities provided by the unlabeled dataset via our proposed UIR loss. In order to 'reject' faces of unknown identities, centers of the known identities are forced to keep enough margin from centers of unknown identities which are assumed to be approximated by the features of their samples. By this means, the discriminativeness of the face representations can be enhanced. Experimental results demonstrate that our approach can provide obvious performance improvement by utilizing the unlabeled data.

No abstract