Tongji University

About: Tongji University is a education organization based out in Shanghai, China. It is known for research contribution in the topics: Population & Adsorption. The organization has 76116 authors who have published 81176 publications receiving 1248911 citations. The organization is also known as: Tongji & Tóngjì Dàxué.

Semi-supervised Medical Image Segmentation through Dual-task Consistency

Abstract: Deep learning-based semi-supervised learning (SSL) algorithms have led to promising results in medical images segmentation and can alleviate doctors' expensive annotations by leveraging unlabeled data. However, most of the existing SSL algorithms in literature tend to regularize the model training by perturbing networks and/or data. Observing that multi/dual-task learning attends to various levels of information which have inherent prediction perturbation, we ask the question in this work: can we explicitly build task-level regularization rather than implicitly constructing networks- and/or data-level perturbation-and-transformation for SSL? To answer this question, we propose a novel dual-task-consistency semi-supervised framework for the first time. Concretely, we use a dual-task deep network that jointly predicts a pixel-wise segmentation map and a geometry-aware level set representation of the target. The level set representation is converted to an approximated segmentation map through a differentiable task transform layer. Simultaneously, we introduce a dual-task consistency regularization between the level set-derived segmentation maps and directly predicted segmentation maps for both labeled and unlabeled data. Extensive experiments on two public datasets show that our method can largely improve the performance by incorporating the unlabeled data. Meanwhile, our framework outperforms the state-of-the-art semi-supervised medical image segmentation methods. Code is available at: this https URL

Aerobic denitrification in permeable Wadden Sea sediments.

Abstract: Permeable or sandy sediments cover the majority of the seafloor on continental shelves worldwide, but little is known about their role in the coastal nitrogen cycle. We investigated the rates and controls of nitrogen loss at a sand flat (Janssand) in the central German Wadden Sea using multiple experimental approaches, including the nitrogen isotope pairing technique in intact core incubations, slurry incubations, a flow-through stirred retention reactor and microsensor measurements. Results indicate that permeable Janssand sediments are characterized by some of the highest potential denitrification rates (> or =0.19 mmol N m(-2) h(-1)) in the marine environment. Moreover, several lines of evidence showed that denitrification occurred under oxic conditions. In intact cores, microsensor measurements showed that the zones of nitrate/nitrite and O(2) consumption overlapped. In slurry incubations conducted with (15)NO(3)(-) enrichment in gas-impermeable bags, denitrification assays revealed that N(2) production occurred at initial O(2) concentrations of up to approximately 90 microM. Initial denitrification rates were not substantially affected by O(2) in surficial (0-4 cm) sediments, whereas rates increased by twofold with O(2) depletion in the at 4-6 cm depth interval. In a well mixed, flow-through stirred retention reactor (FTSRR), (29)N(2) and (30)N(2) were produced and O(2) was consumed simultaneously, as measured online using membrane inlet mass spectrometry. We hypothesize that the observed high denitrification rates in the presence of O(2) may result from the adaptation of denitrifying bacteria to recurrent tidally induced redox oscillations in permeable sediments at Janssand.

Influence of Urban Form on Travel Behaviour in Four Neighbourhoods of Shanghai

Abstract: Since the 1980s, much of the new urban development in China has departed significantly from the traditional pedestrian and bicycle-oriented urban form. This study examines the effects of the urban spatial transformation on travel, based on a travel survey of 1709 individuals from four selected neighbourhoods of Shanghai. Although pedestrian/ cyclist-friendly urban form may not avert the general trend of motorisation growth induced by rising income, it does help to slow down the pace of growth and reduces the need for high levels of motor vehicle ownership. In the pedestrian/cyclist-friendly neighbourhoods, residents travel shorter distances than in other neighbourhoods even though the same modes of travel are used. Pedestrian/cyclist-friendly urban form makes the non-motorised modes feasible options, which is essential to limit automobile dependence. These findings suggest that land use planning and urban design can effectively influence people's mobility demand and travel behaviour towards achieving `gre...