A denoising model based on multi-agent reinforcement learning with data transformation for digital tomosynthesis.

Objective. Denoising models based on the supervised learning have been proposed for medical imaging. However, its clinical availability in digital tomosynthesis (DT) imaging is limited due to the necessity of a large amount of training data for providing acceptable image quality and the difficulty in minimizing a loss. Reinforcement learning (RL) can provide the optimal pollicy, which maximizes a reward, with a small amount of training data for implementing a task. In this study, we presented a denoising model based on the multi-agent RL for DT imaging in order to improve the performance of the machine learning-based denoising model.Approach. The proposed multi-agent RL network consisted of shared sub-network, value sub-network with a reward map convolution (RMC) technique and policy sub-network with a convolutional gated recurrent unit (convGRU). Each sub-network was designed for implementing feature extraction, reward calculation and action execution, respectively. The agents of the proposed network were assigned to each image pixel. The wavelet and Anscombe transformations were applied to DT images for delivering precise noise features during network training. The network training was implemented with the DT images obtained from the three-dimensional digital chest phantoms, which were constructed by using clinical CT images. The performance of the proposed denoising model was evaluated in terms of signal-to-noise ratio (SNR), structural similarity (SSIM) and peak signal-to-noise ratio (PSNR).Main results. Comparing the supervised learning, the proposed denoising model improved the SNRs of the output DT images by 20.64% while maintaining the similar SSIMs and PSNRs. In addition, the SNRs of the output DT images with the wavelet and Anscombe transformations were 25.88 and 42.95% higher than that for the supervised learning, respectively.Significance. The denoising model based on the multi-agent RL can provide high-quality DT images, and the proposed method enables the performance improvement of machine learning-based denoising models.

Multi-bandgap Solar Energy Conversion via Combination of Microalgal Photosynthesis and Spectrally Selective Photovoltaic Cell.

Microalgal photosynthesis is a promising solar energy conversion process to produce high concentration biomass, which can be utilized in the various fields including bioenergy, food resources, and medicine. In this research, we study the optical design rule for microalgal cultivation systems, to efficiently utilize the solar energy and improve the photosynthesis efficiency. First, an organic luminescent dye of 3,6-Bis(4'-(diphenylamino)-1,1'-biphenyl-4-yl)-2,5-dihexyl-2,5-dihydropyrrolo3,4-c pyrrole -1,4-dione (D1) was coated on a photobioreactor (PBR) for microalgal cultivation. Unlike previous reports, there was no enhancement in the biomass productivities under artificial solar illuminations of 0.2 and 0.6 sun. We analyze the limitations and future design principles of the PBRs using photoluminescence under strong illumination. Second, as a multiple-bandgaps-scheme to maximize the conversion efficiency of solar energy, we propose a dual-energy generator that combines microalgal cultivation with spectrally selective photovoltaic cells (PVs). In the proposed system, the blue and green photons, of which high energy is not efficiently utilized in photosynthesis, are absorbed by a large-bandgap PV, generating electricity with a high open-circuit voltage (Voc) in reward for narrowing the absorption spectrum. Then, the unabsorbed red photons are guided into PBR and utilized for photosynthesis with high efficiency. Under an illumination of 7.2 kWh m-2 d-1, we experimentally verified that our dual-energy generator with C60-based PV can simultaneously produce 20.3 g m-2 d-1 of biomass and 220 Wh m-2 d-1 of electricity by utilizing multiple bandgaps in a single system.

Increased biomass and lipid production of Ettlia sp. YC001 by optimized C and N sources in heterotrophic culture.

The culture conditions and media composition for the heterotrophic culture of an axenic strain of Ettlia sp. YC001 were firstly optimized using the Plackett-Burman design (PBD) and response surface methodology (RSM). The strain successfully showed higher productivity in the basal media without any light illumination at 32.2 to 33.3 °C. The PBD results showed that the most effective components for biomass productivity of Ettlia sp. were fructose and yeast extract for sources of C and N, respectively. The RSM results showed an optimal level of 72.2 g/L for fructose and 21.5 g/L for yeast extract, resulting in 46.1 g/L biomass with a lipid content of 13.8% over a course of 9 days. Using a 5 L scaled-up fermentation system for 6 days, the production of biomass and lipids was 7.21 g/L/day and 1.18 g/L/day, respectively. Consequently, heterotrophic cultivation of Ettlia sp. YC001 provided much higher production of biomass and lipids than those of autotrophic cultivation. As further research, the use of substitute substrates instead of fructose and yeast extract should be developed to reduce production costs.

Study of Optical Configurations for Multiple Enhancement of Microalgal Biomass Production.

Microalga is a promising biomass feedstock to restore the global carbon balance and produce sustainable bioenergy. However, the present biomass productivity of microalgae is not high enough to be marketable mainly because of the inefficient utilization of solar energy. Here, we study optical engineering strategies to lead to a breakthrough in the biomass productivity and photosynthesis efficiency of a microalgae cultivation system. Our innovative optical system modelling reveals the theoretical potential (>100 g m-2 day-1) of the biomass productivity and it is used to compare the optical aspects of various photobioreactor designs previously proposed. Based on the optical analysis, the optimized V-shaped configuration experimentally demonstrates an enhancement of biomass productivity from 20.7 m-2 day-1 to 52.0 g m-2 day-1, under the solar-simulating illumination of 7.2 kWh m-2 day-1, through the dilution and trapping of incident energy. The importance of quantitative optical study for microalgal photosynthesis is clearly exhibited with practical demonstration of the doubled light utilization efficiencies.

Enhancement of lipid productivity by adopting multi-stage continuous cultivation strategy in Nannochloropsis gaditana.

In the present study, a novel process-based cultivation system was designed to improve lipid productivity of Nannochloropsis gaditana, an oleaginous microalga that has high potential for biofuel production. Specifically, four flat-panel photobioreactors were connected in series, and this system was subjected to continuous chemostat cultivation by feeding fresh medium to the first reactor at dilution rates of 0.028 and 0.056day-1, which were determined based on Monod kinetics. The results show that the serially connected photobioreactor system achieved 20.0% higher biomass productivity and 46.1% higher fatty acid methyl ester (FAME) productivity than a conventional single photobioreactor with equivalent dilution rate. These results suggest that a process-based approach using serially connected photobioreactors for microalgal cultivation can improve the productivity of lipids that can be used for biofuel production.

Simultaneous treatment of food-waste recycling wastewater and cultivation of Tetraselmis suecica for biodiesel production.

There is an increasing interest in the use of cultivated microalgae to simultaneously produce biodiesel and remove nutrients from various wastewaters. For this purpose, Tetraselmis suecica was cultivated in flasks and fermenters using diluted food-waste recycling wastewater (FRW). The effect of FRW dilution on T. suecica growth and nutrient removal was initially tested in flasks. The maximal microalgal concentration after 14 days was in medium with a twofold dilution (28.3 × 10(6) cells/mL) and a fivefold dilution (25.5 × 10(6) cells/mL). When fivefold diluted medium was used in fermenters, the final dry cell weight of T. suecica was 2.0 g/L. The removal efficiencies of ammonium and phosphate in the fermenters were 99.0 and 52.3%, respectively. In comparison with the results of previous studies, the growth data of T. suecica in the FRW medium indicate that microalgal cultivation system incorporating removal of nutrients in FRW is feasible at the field level.

A comprehensive study on algal-bacterial communities shift during thiocyanate degradation in a microalga-mediated process.

Changes in algal and bacterial communities during thiocyanate (SCN(-)) decomposition in a microalga-mediated process were studied. Pyrosequencing indicated that Thiobacillus bacteria and Micractinium algae predominated during SCN(-) hydrolysis, even after its complete degradation. Principal components analysis and evenness profiles (based on the Pareto-Lorenz curve) suggested that the changes in the bacterial communities were driven by nitrogen and sulfur oxidation, pH changes, and photoautotrophic conditions. The populations of predominant microalgae remained relatively stable during SCN(-) hydrolysis, but the proportion of bacteria - especially nitrifying bacteria - fluctuated. Thus, the initial microalgal population may be crucial in determining which microorganisms dominate when the preferred nitrogen source becomes limited. The results also demonstrated that microalgae and SCN(-)-hydrolyzing bacteria can coexist, that microalgae can be effectively used with these bacteria to completely treat SCN(-), and that the structure of the algal-bacterial community is more stable than the community of nitrifying bacteria alone during SCN(-) degradation.

Use of a triiodide resin for isolation of axenic cultures of microalgal Nannochloropsis gaditana.

Triiodide resin (TR) was used to generate axenic cultures of microalgae by employing the antibacterial capability of triiodide. A Nannochloropsis gaditana culture contaminated with bacteria was passed through a column filled with TR using the gravity flow. Based on analyses of flow cytometry and vital staining using a fluorescent dye SYTOX Green, three cycles of TR treatments remarkably reduced the number of viable bacteria but had little effects on the microalgae. This novel approach is a simple, rapid, and cost-effective method that can be used to isolate axenic cultures of microalgae.

A simple method for decomposition of peracetic acid in a microalgal cultivation system.

A cost-efficient process devoid of several washing steps was developed, which is related to direct cultivation following the decomposition of the sterilizer. Peracetic acid (PAA) is known to be an efficient antimicrobial agent due to its high oxidizing potential. Sterilization by 2 mM PAA demands at least 1 h incubation time for an effective disinfection. Direct degradation of PAA was demonstrated by utilizing components in conventional algal medium. Consequently, ferric ion and pH buffer (HEPES) showed a synergetic effect for the decomposition of PAA within 6 h. On the contrary, NaNO3, one of the main components in algal media, inhibits the decomposition of PAA. The improved growth of Chlorella vulgaris and Synechocystis PCC6803 was observed in the prepared BG11 by decomposition of PAA. This process involving sterilization and decomposition of PAA should help cost-efficient management of photobioreactors in a large scale for the production of value-added products and biofuels from microalgal biomass.

Cultivation of Chlorella protothecoides in anaerobically treated brewery wastewater for cost-effective biodiesel production.

The use of wastewater has been investigated to overcome the economic challenge involved with a production of microalgae-based biodiesel. In this study, to achieve economical biodiesel production along with effective wastewater treatment at the same time, anaerobically treated brewery wastewater (ABWW) was utilized as a low-cost nutrient source, in the cultivation of Chlorella protothecoides. About 96 and 90 % of total nitrogen and phosphorus in ABWW were removed, respectively, while C. protothecoides was accumulating 1.88 g L(-1) of biomass. The C. protothecoides grown in ABWW showed increases in cell size and cell aggregation, resulting in a near 80 % enhanced harvesting efficiency within 20 min, as compared with only 4 % in BG-11. In addition, the total fatty acid content of the C. protothecoides grown in ABWW increased by 1.84-fold (35.94 ± 1.54 % of its dry cell weight), relative to that of BG-11.

Effect of monochromatic illumination on lipid accumulation of Nannochloropsis gaditana under continuous cultivation.

Although nitrogen starvation is frequently used to increase lipid contents in microalgae, it has a negative effect on cellular growth. Since light supply is essential for photosynthetic organisms, the effects of cultivation under monochromatic illumination on the growth and lipid contents of Nannochloropsis gaditana were assessed. Continuous cultivation under blue and red light conditions improved the productivity and physical properties for biodiesel from this microalga. FAME yield was twofold higher under red light than under normal white light (21.12% vs 11.35%), with no significant difference in growth rates. Blue and red light increased photosynthetic oxygen evolution, carbon fixation and nutrient uptake. In total, more significant physiological changes were observed under red than under blue light. These results show that red light illumination may be useful for enhancing lipid production by N. gaditana, with the increased photosynthetic reducing equivalents induced by red light which could be deposited as lipids and carbohydrates.