Enhanced fabrication of size-controllable chitosan-genipin nanoparticles using orifice-induced hydrodynamic cavitation: Process optimization and performance evaluation.

Chitosan nanoparticles (NPs) possess great potential in biomedical fields. Orifice-induced hydrodynamic cavitation (HC) has been used for the enhancement of fabrication of size-controllable genipin-crosslinked chitosan (chitosan-genipin) NPs based on the emulsion cross-linking (ECLK). Experiments have been performed using various plate geometries, chitosan molecular weight and under different operational parameters such as inlet pressure (1-3.5 bar), outlet pressure (0-1.5 bar) and cross-linking temperature (40-70 °C). Orifice plate geometry was a crucial factor affecting the properties of NPs, and the optimized geometry of orifice plate was with single hole of 3.0 mm diameter. The size of NPs with polydispersity index of 0.359 was 312.6 nm at an optimized inlet pressure of 3.0 bar, and the maximum production yield reached 84.82 %. Chitosan with too high or too low initial molecular weight (e.g., chitosan oligosaccharide) was not applicable for producing ultra-fine and narrow-distributed NPs. There existed a non-linear monotonically-increasing relationship between cavitation number (Cv) and chitosan NP size. Scanning electron microscopy (SEM) test indicated that the prepared NPs were discrete with spherical shape. The study demonstrated the superiority of HC in reducing particle size and size distribution of NPs, and the energy efficiency of orifice type HC-processed ECLK was two orders of magnitude than that of ultrasonic horn or high shear homogenization-processed ECLK. In vitro drug-release studies showed that the fabricated NPs had great potential as a drug delivery system. The observations of this study can offer strong support for HC to enhance the fabrication of size-controllable chitosan-genipin NPs.

Sequential catalytic lignin valorization and bioethanol production: an integrated biorefinery strategy.

BACKGROUND: The effective valorization of lignin and carbohydrates in lignocellulose matrix under the concept of biorefinery is a primary strategy to produce sustainable chemicals and fuels. Based on the reductive catalytic fractionation (RCF), lignin in lignocelluloses can be depolymerized into viscous oils, while the highly delignified pulps with high polysaccharides retention can be transformed into various chemicals. RESULTS: A biorefinery paradigm for sequentially valorization of the main components in poplar sawdust was constructed. In this process, the well-defined low-molecular-weight phenols and bioethanol were co-generated by tandem chemo-catalysis in the RCF stage and bio-catalysis in fermentation stage. In the RCF stage, hydrogen transfer reactions were conducted in one-pot process using Raney Ni as catalyst, while the isopropanol (2-PrOH) in the initial liquor was served as a hydrogen donor and the solvent for lignin dissolution. Results indicated the proportion of the 2-PrOH in the initial liquor of RCF influenced the chemical constitution and yield of the lignin oil, which also affected the characteristics of the pulps and the following bioethanol production. A 67.48 ± 0.44% delignification with 20.65 ± 0.31% of monolignols yield were realized when the 2-PrOH:H2O ratio in initial liquor was 7:3 (6.67 wt% of the catalyst loading, 200 °C for 3 h). The RCF pulp had higher carbohydrates retention (57.96 ± 2.78 wt%), which was converted to 21.61 ± 0.62 g/L of bioethanol with a yield of 0.429 ± 0.010 g/g in fermentation using an engineered S. cerevisiae strain. Based on the mass balance analysis, 104.4 g of ethanol and 206.5 g of lignin oil can be produced from 1000 g of the raw poplar sawdust. CONCLUSIONS: The main chemical components in poplar sawdust can be effectively transformed into lignin oil and bioethanol. The attractive results from the biorefinery process exhibit great promise for the production of valuable biofuels and chemicals from abundant lignocellulosic materials.

Structured Label Inference for Visual Understanding.

Visual data such as images and videos contain a rich source of structured semantic labels as well as a wide range of interacting components. Visual content could be assigned with fine-grained labels describing major components, coarse-grained labels depicting high level abstractions, or a set of labels revealing attributes. Such categorization over different, interacting layers of labels evinces the potential for a graph-based encoding of label information. In this paper, we exploit this rich structure for performing graph-based inference in label space for a number of tasks: multi-label image and video classification and action detection in untrimmed videos. We consider the use of the Bidirectional Inference Neural Network (BINN) and Structured Inference Neural Network (SINN) for performing graph-based inference in label space and propose a Long Short-Term Memory (LSTM) based extension for exploiting activity progression on untrimmed videos. The methods were evaluated on (i) the Animal with Attributes (AwA), Scene Understanding (SUN) and NUS-WIDE datasets for multi-label image classification, (ii) the first two releases of the YouTube-8M large scale dataset for multi-label video classification, and (iii) the THUMOS'14 and MultiTHUMOS video datasets for action detection. Our results demonstrate the effectiveness of structured label inference in these challenging tasks, achieving significant improvements against baselines.

A Subdivision-Based Representation for Vector Image Editing.

Vector graphics has been employed in a wide variety of applications due to its scalability and editability. Editability is a high priority for artists and designers who wish to produce vector-based graphical content with user interaction. In this paper, we introduce a new vector image representation based on piecewise smooth subdivision surfaces, which is a simple, unified and flexible framework that supports a variety of operations, including shape editing, color editing, image stylization, and vector image processing. These operations effectively create novel vector graphics by reusing and altering existing image vectorization results. Because image vectorization yields an abstraction of the original raster image, controlling the level of detail of this abstraction is highly desirable. To this end, we design a feature-oriented vector image pyramid that offers multiple levels of abstraction simultaneously. Our new vector image representation can be rasterized efficiently using GPU-accelerated subdivision. Experiments indicate that our vector image representation achieves high visual quality and better supports editing operations than existing representations.