Learning A Single Network for Robust Medical Image Segmentation with Noisy Labels.

Robust segmenting with noisy labels is an important problem in medical imaging due to the difficulty of acquiring high-quality annotations. Despite the enormous success of recent developments, these developments still require multiple networks to construct their frameworks and focus on limited application scenarios, which leads to inflexibility in practical applications. They also do not explicitly consider the coarse boundary label problem, which results in sub-optimal results. To overcome these challenges, we propose a novel Simultaneous Edge Alignment and Memory-Assisted Learning (SEAMAL) framework for noisy-label robust segmentation. It achieves single-network robust learning, which is applicable for both 2D and 3D segmentation, in both Set-HQ-knowable and Set-HQ-agnostic scenarios. Specifically, to achieve single-model noise robustness, we design a Memory-assisted Selection and Correction module (MSC) that utilizes predictive history consistency from the Prediction Memory Bank to distinguish between reliable and non-reliable labels pixel-wisely, and that updates the reliable ones at the superpixel level. To overcome the coarse boundary label problem, which is common in practice, and to better utilize shape-relevant information at the boundary, we propose an Edge Detection Branch (EDB) that explicitly learns the boundary via an edge detection layer with only slight additional computational cost, and we improve the sharpness and precision of the boundary with a thinning loss. Extensive experiments verify that SEAMAL outperforms previous works significantly.

Integrated municipal wastewater treatment and lipid accumulation by a self-flocculating/floating microalga Limnothrix sp.

The present study evaluated the growth, self-flocculation, lipid content, and pollutants removal by Limnothrix sp. BASMWW-9 isolated from municipal wastewater treatment system and cultivated in municipal wastewater. The biomass yield and lipid content after 6 days of cultivation were 1.07 g dw/L and 27.34 %dw, respectively. In addition, its self-flocculating ability reached up to 90 % after harvesting time of 180 min. Moreover, COD,NH3-N, TN, and TP removalefficiencies were 71.65 %, 81.89 %, 74.64 %, and 80.16 %, respectively. The self-flocculation performance of Limnothrix sp. was greatly associated to its morphology and production of extracellular polymeric substances (EPS), with significant positive impact of the high calcium and magnesium content in municipal wastewater. Interestingly, blue light irradiation during harvest enhanced the aggregation and floc formation as a floating biomat, which was attributed to enhanced polysaccharides production. This study provides innovative harvest method for Limnothrix sp. BASMWW-9 cultivated in wastewater using blue light for enhanced lipid recovery.

Point-Guided Contrastive Learning for Monocular 3-D Object Detection.

3-D object detection is a fundamental task in the context of autonomous driving. In the literature, cheap monocular image-based methods show a significant performance drop compared to the expensive LiDAR and stereo-images-based algorithms. In this article, we aim to close this performance gap by bridging the representation capability between 2-D and 3-D domains. We propose a novel monocular 3-D object detection model using self-supervised learning and auxiliary learning, resorting to mimicking the representations over 3-D point clouds. Specifically, given a 2-D region proposal and the corresponding instance point cloud, we supervise the feature activation from our image-based convolution network to mimic the latent feature of a point-based neural network at the training stage. While state-of-the-art (SOTA) monocular 3-D detection algorithms typically convert images to pseudo-LiDAR with depth estimation and regress 3-D detection with LiDAR-based methods, our approach seeks the power of the 2-D neural network straightforwardly and essentially enhances the 2-D module capability with latent spatial-aware representations by contrastive learning. We empirically validate the performance improvement from the feature mimicking the KITTI and ApolloScape datasets and achieve the SOTA performance on the KITTI and ApolloScape leaderboard.

Robust Point Cloud Segmentation With Noisy Annotations.

Point cloud segmentation is a fundamental task in 3D. Despite recent progress on point cloud segmentation with the power of deep networks, current learning methods based on the clean label assumptions may fail with noisy labels. Yet, class labels are often mislabeled at both instance-level and boundary-level in real-world datasets. In this work, we take the lead in solving the instance-level label noise by proposing a Point Noise-Adaptive Learning (PNAL) framework. Compared to noise-robust methods on image tasks, our framework is noise-rate blind, to cope with the spatially variant noise rate specific to point clouds. Specifically, we propose a point-wise confidence selection to obtain reliable labels from the historical predictions of each point. A cluster-wise label correction is proposed with a voting strategy to generate the best possible label by considering the neighbor correlations. To handle boundary-level label noise, we also propose a variant "PNAL-boundary " with a progressive boundary label cleaning strategy. Extensive experiments demonstrate its effectiveness on both synthetic and real-world noisy datasets. Even with 60% symmetric noise and high-level boundary noise, our framework significantly outperforms its baselines, and is comparable to the upper bound trained on completely clean data. Moreover, we cleaned the popular real-world dataset ScanNetV2 for rigorous experiment. Our code and data is available at https://github.com/pleaseconnectwifi/PNAL.

3D Question Answering.

Visual question answering (VQA) has experienced tremendous progress in recent years. However, most efforts have only focused on 2D image question-answering tasks. In this paper, we extend VQA to its 3D counterpart, 3D question answering (3DQA), which can facilitate a machine's perception of 3D real-world scenarios. Unlike 2D image VQA, 3DQA takes the color point cloud as input and requires both appearance and 3D geometrical comprehension to answer the 3D-related questions. To this end, we propose a novel transformer-based 3DQA framework "3DQA-TR", which consists of two encoders to exploit the appearance and geometry information, respectively. Finally, the multi-modal information about the appearance, geometry, and linguistic question can attend to each other via a 3D-linguistic Bert to predict the target answers. To verify the effectiveness of our proposed 3DQA framework, we further develop the first 3DQA dataset "ScanQA", which builds on the ScanNet dataset and contains over 10 K question-answer pairs for 806 scenes. To the best of our knowledge, ScanQA is the first large-scale dataset with natural-language questions and free-form answers in 3D environments that is fully human-annotated. We also use several visualizations and experiments to investigate the astonishing diversity of the collected questions and the significant differences between this task from 2D VQA and 3D captioning. Extensive experiments on this dataset demonstrate the obvious superiority of our proposed 3DQA framework over state-of-the-art VQA frameworks and the effectiveness of our major designs. Our code and dataset will be made publicly available to facilitate research in this direction. The code and data are available at http://shuquanye.com/3DQA\_website/.

Efficient simultaneous removal of tetracycline hydrochloride and Cr(VI) through photothermal-assisted photocatalytic-Fenton-like processes with CuOx/γ-Al2O3.

Advanced oxidation processes (AOPs) based on photo-Fenton reaction has been widely used in refractory organic wastewater treatment. However, the mineralization rate and H2O2 utilization in AOPs is relatively low. Herein, a photothermal-assisted photocatalytic-Fenton-like process with CuOx/γ-Al2O3 catalyst was designed to solve the above issues. The utilization rate of H2O2 and mineralization rate of tetracycline hydrochloride (TC) in the 3%-CuOx/γ-Al2O3/H2O2/full spectrum light system were significantly increased to 72.0% and 74.3%, respectively, which were ascribed to the synergy of photothermal effect and photocatalytic-Fenton-like reaction. During the simultaneous removal of TC and Cr(VI), the reaction efficiency of 3%-CuOx/γ-Al2O3/H2O2/light system was much higher than that of 3%-CuOx/γ-Al2O3/H2O2 system. In addition, the 3%-CuOx/γ-Al2O3/H2O2/light system was not only active in wider pH range (3-9), but also effective in various organic pollutants and Cr(VI) coexisted solutions as well as different water conditions. More importantly, a possible photothermal-assisted photocatalytic-Fenton-like reaction mechanism for the simultaneous removal of TC and Cr(VI) in the CuOx/γ-Al2O3/H2O2/light system was proposed. This work may pave a new way for efficient simultaneous removal of refractory organic pollutants and heavy metals in wastewater.

Meta-PU: An Arbitrary-Scale Upsampling Network for Point Cloud.

Point cloud upsampling is vital for the quality of the mesh in three-dimensional reconstruction. Recent research on point cloud upsampling has achieved great success due to the development of deep learning. However, the existing methods regard point cloud upsampling of different scale factors as independent tasks. Thus, the methods need to train a specific model for each scale factor, which is both inefficient and impractical for storage and computation in real applications. To address this limitation, in this article, we propose a novel method called "Meta-PU" to first support point cloud upsampling of arbitrary scale factors with a single model. In the Meta-PU method, besides the backbone network consisting of residual graph convolution (RGC) blocks, a meta-subnetwork is learned to adjust the weights of the RGC blocks dynamically, and a farthest sampling block is adopted to sample different numbers of points. Together, these two blocks enable our Meta-PU to continuously upsample the point cloud with arbitrary scale factors by using only a single model. In addition, the experiments reveal that training on multiple scales simultaneously is beneficial to each other. Thus, Meta-PU even outperforms the existing methods trained for a specific scale factor only.

Visibility-Aware Point-Based Multi-View Stereo Network.

We introduce VA-Point-MVSNet, a novel visibility-aware point-based deep framework for multi-view stereo (MVS). Distinct from existing cost volume approaches, our method directly processes the target scene as point clouds. More specifically, our method predicts the depth in a coarse-to-fine manner. We first generate a coarse depth map, convert it into a point cloud and refine the point cloud iteratively by estimating the residual between the depth of the current iteration and that of the ground truth. Our network leverages 3D geometry priors and 2D texture information jointly and effectively by fusing them into a feature-augmented point cloud, and processes the point cloud to estimate the 3D flow for each point. This point-based architecture allows higher accuracy, more computational efficiency and more flexibility than cost-volume-based counterparts. Furthermore, our visibility-aware multi-view feature aggregation allows the network to aggregate multi-view appearance cues while taking into account visibility. Experimental results show that our approach achieves a significant improvement in reconstruction quality compared with state-of-the-art methods on the DTU and the Tanks and Temples dataset. The code of VA-Point-MVSNet proposed in this work will be released at https://github.com/callmeray/PointMVSNet.

Sequential 'acid leaching-anion exchange-aerobic composting' process for recycle of municipal sludge.

Municipal sludge disposal and recycle has become a prominent research theme. In this study, a sequential process for integral treatment of municipal sludge was primarily presented, combining acid leaching, anion exchange and aerobic composting. The aim of the process was to remove chromium (Cr) from the sludge and reuse the sludge as manure. Firstly, Cr was removed from municipal sludge via the acid leaching process; the removal rate was up to 57.43%. Then, ion exchange resin was used to remove Cr from leachate; the removal rate reached 95%. Aluminum sheet was used to replace the Cr from eluent; the replacement rate was 63.3%. The aerobic composting process could be successfully warmed up to above 55 °C and lasted for 4 days; the seed germination index reached 68.3%. After the composting process, the residual Cr in sludge mainly existed at a more stable residual state and organic binding state. Overall, this novel sequential process serves as a potential high-efficiency, green, low-energy way for municipal sludge recycle.

Microalgae harvesting by magnetic flocculation for biodiesel production: current status and potential.

With the increasing demand for energy, microalgae, as one of the promising feedstocks of biodiesel, has raised great awareness. Because of its small size, similar density to water and electrical stability, harvesting methods of microalgae that have low energy consumption and that are highly efficient, easy to large-scale and environmentally friendly have become a bottleneck restricting development of the whole process. Among the numerous possible harvesting methods, magnetic flocculation has the advantages of simple operation, fast separation and energy saving and thus is considered as a promising novel harvesting method. In this review, we have summarized the updated status and application potential of magnetic flocculation, including the principle of magnetic flocculation, magnetic flocculating materials, flocculating efficiency and its effect on downstream process. The major challenges such as magnetic materials recovery, large-scale magnetic flocculation device design, and magnetic flocculation costs are also discussed.

Enhanced lipid extraction from the biodiesel-producing microalga Chlorella pyrenoidosa cultivated in municipal wastewater via Daphnia ingestion and digestion.

Herein, a novel innovative lipid co-extraction strategy using the biodiesel-producing microalga Chlorella pyrenoidosa and planktonic cladoceran Daphnia was proposed. Co-extraction occurred as Daphnia ingested and digested microalgal cells in a pre-treatment process; thereafter, lipids from these organisms were extracted. Composition of fatty acids from C. pyrenoidosa and Daphnia were appropriate as potential biodiesel feedstocks. Daphnia had different absorption and conversion capacities of various fatty acids from C. pyrenoidosa, which showed potential for improving biodiesel characteristics. Linoleic acid (LA, C18:2n-6) and alpha-linolenic acid (ALA, C18:3n-3) were absorbed significantly into the body of Daphnia. The optimal lipid extraction and fatty acid methyl esters (FAMEs) recovery rates were up to 41.08% and 12.35%, respectively, which were greater than that of the traditional lipid extraction method due to the rich oil content of Daphnia. Overall, this lipid co-extraction process serves a potential Daphnia utilization as an economical, green, low-energy way for microalgae biodiesel production.

Enhanced lipid production by Chlorella pyrenoidosa through magnetic field pretreatment of wastewater and treatment of microalgae-wastewater culture solution: Magnetic field treatment modes and conditions.

The cultivation of microalgae in municipal wastewater not only purifies the wastewater but also transforms nutrients into biomass that contains high-value lipids. In this study, conventional static bottom-magnetic field (bottom-MF) equipment and cost-effective bypass-magnetic field (bypass-MF) equipment were designed and independently coupled with a microalgae-wastewater system in different positions to evaluate the effect of magnetic field (MF) on microalgae biomass production and lipid accumulation. When the MF equipment was applied in the wastewater pretreatment unit, the bottom-MF pretreatment mode exhibited a more beneficial effect on subsequent biomass and lipid accumulation. However, when the MF equipment was applied in the microalgae-wastewater culture unit, there was no significant difference between the bottom-MF and bypass-MF modes. The results of the orthogonal experiment suggested the optimum conditions for lipid production were wastewater pretreatment by bottom-MF at 5000 Gs for 1 h, followed by microalgae-wastewater culture treatment by bypass-MF at 5000 Gs for 3 h.

Screening and mutagenesis of high-efficient degrading bacteria of linear alkylbenzene sulfonates.

As a widely used detergent, anionic surfactant linear alkylbenzene sulfonates (LAS) is a common toxic pollutant in wastewater. In this study, Pseudomonas sp. strain H6 was isolated from activated sludge and municipal wastewater, which had good degradation effect on LAS. The results showed that strain H6 could grow with LAS as the sole carbon source. When the concentration of LAS was less than 100 mg/L, strain H6 could degrade more than 80% of the LAS within 24 h. Meanwhile, the growth of strain H6 increased with the increase of LAS concentration, reaching the maximum growth at the presence of 100 mg/L LAS. When the concentration of LAS was over 100 mg/L, strain H6's cell growth and degradation of LAS showed a downward trend due to the strong toxicity of LAS, and the degradation rate of LAS almost tended to zero with 500 mg/L LAS. Further mutagenesis analysis of strain H6 showed that positive mutation occurred under ultraviolet and nitrite mutagenesis with using ampicillin to increase the screening pressure, and the degradation rate of LAS was 44.91% higher than that of original strain.

Rapid enrichment and ammonia oxidation performance of ammonia-oxidizing archaea from an urban polluted river of China.

Ammonia oxidation is the rate-limiting step in nitrification process and dominated by ammonia-oxidizing bacteria (AOB) and ammonia-oxidizing archaea (AOA). In the present study, a highly enriched culture of AOA was obtained from urban polluted water in Shahe River, Shenzhen, China. The optimum growth conditions were identified by orthogonal analysis as 37 °C, with pH 7.0 and initial ammonia concentration of 1.0 mM. Under these conditions, the highest abundance of AOA was obtained as 4.6 × 107 copies/ng DNA. Growth of AOA in polluted river water showed significant reduction in ammonia concentration in AOA-enriched cultures without antibiotics after 10 days of incubation, while synchronous increase in nitrate concentration was up to 12.7 mg/L. However, AOA-enriched by antibiotic showed insignificant changes in ammonia or nitrite concentration. This study showed that AOB play an important role in ammonia oxidation of polluted river water, and AOA alone showed insignificant changes in ammonia or nitrite concentrations. Therefore, the ammonia oxidation performance of natural water could not be improved by adding high concentration AOA bacterial liquid.

Efficient harvesting of Chlorella pyrenoidosa and Scenedesmus obliquus cultivated in urban sewage by magnetic flocculation using nano-Fe3O4 coated with polyethyleneimine.

In this work, a novel flocculation process by using nano-Fe3O4 coated with polyethyleneimine (Fe3O4@PEI) as magnetic seeds was developed to harvest the microalgae cultivated in urban sewage. Experiment results indicated that the harvest efficiency of Chlorella pyrenoidosa (0.5 g/L) was 98.92 ±â€¯0.41% under the optimal conditions of Fe3O4@PEI dose of 20 mL/L, flocculation time of 20 min, and stirring speed of 800 rpm (3 min), while that of Scenedesmus obliquus (0.4 g/L) was 98.45 ±â€¯0.35% under a Fe3O4@PEI dose of 16 mL/L, flocculation time of 15 min, and stirring speed of 730 rpm (3 min). Moreover, the process did not reduce the lipid content of microalgae and quality of biodiesel. After microalgae harvest, Fe3O4@PEI could be recovered by ultrasonication, re-wrapped with polyethyleneimine and reused to reduce operational cost.

Ozone disinfection of chlorine-resistant bacteria in drinking water.

The wide application of chlorine disinfectant for drinking water treatment has led to the appearance of chlorine-resistant bacteria, which pose a severe threat to public health. This study was performed to explore the physiological-biochemical characteristics and environmental influence (pH, temperature, and turbidity) of seven strains of chlorine-resistant bacteria isolated from drinking water. Ozone disinfection was used to investigate the inactivation effect of bacteria and spores. The DNA concentration and cell surface structure variations of typical chlorine-resistant spores (Bacillus cereus spores) were also analysed by real-time qPCR, flow cytometry, and scanning electron microscopy to determine their inactivation mechanisms. The ozone resistance of bacteria (Aeromonas jandaei < Vogesella perlucida < Pelomonas < Bacillus cereus < Aeromonas sobria) was lower than that of spores (Bacillus alvei < Lysinibacillus fusiformis < Bacillus cereus) at an ozone concentration of 1.5 mg/L. More than 99.9% of Bacillus cereus spores were inactivated by increasing ozone concentration and treatment duration. Moreover, the DNA content of Bacillus cereus spores decreased sharply, but approximately 1/4 of the target genes remained. The spore structure exhibited shrinkage and folding after ozone treatment. Both cell structures and gene fragments were damaged by ozone disinfection. These results showed that ozone disinfection is a promising method for inactivating chlorine-resistant bacteria and spores in drinking water.

Subcritical n-hexane/isopropanol extraction of lipid from wet microalgal pastes of Scenedesmus obliquus.

Subcritical co-solvents of n-hexane/isopropanol were primarily utilized to extract lipid from wet microalgal pastes of Scenedesmus obliquus. The effects of key operational parameters were investigated, and the optimal parameters were obtained: solvent ratio of n-hexane to isopropanol was 3:2 (V:V), phase ratio of co-solvents to microalgal biomass was 35:1 (mL:g), reactor stirring speed was 900 rpm, extraction time was 60 min. Additional pretreatment with acid, ultrasonic and microwave as well as enhanced subcritical pressure/heating treatments were also applied to further study their effects on lipid extraction. The results showed that the lipid recovery rate with acid pretreatment was 8.6 and 6.2% higher than ultrasonic and microwave pretreatment; the optimum enhanced subcritical condition was 55 °C with atmospheric pressure. Under optimal operating conditions, the lipid and FAME yield were 13.5 and 7.2%, which was 82.6 and 135.1% higher than the traditional method. The results indicated that the subcritical n-hexane/isopropanol extraction process had promising application potential.

Enhancing dewaterability of waste activated sludge by combined oxidative conditioning process with zero-valent iron and peroxymonosulfate.

The enhancement of sludge dewaterability is of great importance for facilitating the sludge disposal during the operation of wastewater treatment plants. In this study, a novel oxidative conditioning approach was applied to enhance the dewaterability of waste activated sludge by the combination of zero-valent iron (ZVI) and peroxymonosulfate (PMS). It was found that the dewaterability of sludge was significantly improved after the addition of ZVI (0-4 g/g TSS) (TSS: total suspended solids) and PMS (0-1 g/g TSS). The optimal addition amount of ZVI and PMS was 0.25 g/g TSS and 0.1 g/g TSS, respectively, under which the capillary suction time of the sludge was reduced by approximately 50%. The decomposition of sludge flocs could contribute to the improved sludge dewaterability. Economic analysis demonstrated that the proposed conditioning process with ZVI and PMS was more economical than the ZVI + peroxydisulfate and the traditional Fenton conditioning processes.

Improving of lipid productivity of the oleaginous microalgae Chlorella pyrenoidosa via atmospheric and room temperature plasma (ARTP).

In this study, an efficient screening program was established with ARTP. Five strains from oleaginous microalgae Chlorella pyrenoidosa were screened from mutant library after mutagenizing by ARTP. Among them, the optimal mutant strain was named as II-H6. In the BG11 medium, the OD680 of II-H6 in stationary phase were increased by 32.08% than the original strain. Meanwhile, compared with the original strain, the dry weight and lipid productivity of II-H6 were increased by 22.07% and 16.85%, respectively. II-H6 showed a good genetic stability in BG11 medium and the optimum growth temperature and pH were 33°C and 9.0. 18S gene fragment length of II-H6 strain were 1886bp. Analysis of the gene fragment showed that the II-H6 strain had a close relationship to the original strain, and it belonged to the mutation within the genus Chlorella.

[Selection of Microalgae for Biofuel Using Municipal Wastewater as a Resource].

Cultivating microalgae using municipal wastewater can achieve not only treatment of the wastewater but also recovery of algae for use as a biofuel energy source. Wastewater provides necessary nutrients, such as nitrogen and phosphorus, and water for microalgal growth. Because of the complexity of components of municipal wastewater, variety of adaptability, and tolerance to wastewater of different microalgal species, it is necessary to select a suitable microalgal species with high performance in lipid production and identify proper pretreatment of the wastewater to achieve high lipid production using municipal wastewater for algal biofuel production. Based on microalgal growth, lipid production, and clean-up performance of wastewater, we selected Scenedesmus obliquus wild strain and Chlorella pyrenoidosa mutant by ion beam implantation from a test group for the biofuel production. Laboratory test results showed that S. obliquus wild strain and C. pyrenoidosa mutant had respective lipid productions of 0.43 g·L-1 and 0.33 g·L-1, with more C16-C18 fatty acids, which were suitable for biodiesel production. The pollutant removals from the wastewater by S. obliquus wild strain and C. pyrenoidosa mutant were COD, 86.4% vs. 81.8%; NH4+-N, 100.0% vs. 100.0%; TN, 94.3% vs. 94.9%; and TP, 93.4% vs. 94.2% respectively. The two different microalgal strains required different pretreatments. After removal of large particles, the raw wastewater could be directly used for the cultivation of S. obliquus wild strain. To grow C. pyrenoidosa mutant with municipal wastewater, pretreatment procedures including precipitation followed by filtration should be employed.