Due to the rapid development of the generative adversarial networks (GANs) and convolution neural networks (CNN), increasing attention is being paid to face synthesis. In this paper, we address the new and challenging task of facial sketch-to-image synthesis with multiple controllable attributes. To achieve this goal, first, we propose a new attribute classification loss to ensure that the synthesized face image with the facial attributes, which the users desire to have. Second, we employ the reconstruction loss to synthesize the facial texture and structure information. Third, the adversarial loss is used to encourage visual authenticity. By incorporating above losses into a unified framework, our proposed method not only can achieve high-quality sketch-to-image translation, but also allow the users control the facial attributes of synthesized image. Extensive experiments show that user-provided facial attribute information effectively controls the process of facial sketch-to-image translation.

No abstract

This case study described a successful application of the quality by design (QbD) principles to a coupling process development of insulin degludec. Failure mode effects analysis (FMEA) risk analysis was first used to recognize critical process parameters (CPPs). Five CPPs, including coupling temperature (Temp), pH of desB30 solution (pH), reaction time (Time), desB30 concentration (Conc), and molar equivalent of ester per mole of desB30 insulin (MolE), were then investigated using a fractional factorial design. The curvature effect was found significant, indicating the requirement of second-order models. Afterwards, a central composite design was built with an augmented star and center points study. Regression models were developed for the CPPs to predict the purity and yield of predegludec using above experimental data. The R and adjusted R were higher than 96 and 93% for the two models respectively. The Q values were more than 80% indicating a good predictive ability of models. MolE was found to be the most significant factor affecting both yield and purity of predegludec. Temp, pH, and Conc were also significant for predegludec purity, while Time appeared to remarkably influence the yield model. The multi-dimensional design space and normal operating region (NOR) with a robust setpoint were determined using a probability-based Monte-Carlo simulation method. The verified experimental results showed that the design space was reliable and effective. This study enriches the understanding of acetylation process and is instructional to other complicated operations in biopharmaceutical engineering.