Uncertainty-aware Active Domain Adaptive Salient Object Detection.

Due to the advancement of deep learning, the performance of salient object detection (SOD) has been significantly improved. However, deep learning-based techniques require a sizable amount of pixel-wise annotations. To relieve the burden of data annotation, a variety of deep weakly-supervised and unsupervised SOD methods have been proposed, yet the performance gap between them and fully supervised methods remains significant. In this paper, we propose a novel, cost-efficient salient object detection framework, which can adapt models from synthetic data to real-world data with the help of a limited number of actively selected annotations. Specifically, we first construct a synthetic SOD dataset by copying and pasting foreground objects into pure background images. With the masks of foreground objects taken as the ground-truth saliency maps, this dataset can be used for training the SOD model initially. However, due to the large domain gap between synthetic images and real-world images, the performance of the initially trained model on the real-world images is deficient. To transfer the model from the synthetic dataset to the real-world datasets, we further design an uncertainty-aware active domain adaptive algorithm to generate labels for the real-world target images. The prediction variances against data augmentations are utilized to calculate the superpixel-level uncertainty values. For those superpixels with relatively low uncertainty, we directly generate pseudo labels according to the network predictions. Meanwhile, we select a few superpixels with high uncertainty scores and assign labels to them manually. This labeling strategy is capable of generating high-quality labels without incurring too much annotation cost. Experimental results on six benchmark SOD datasets demonstrate that our method outperforms the existing state-of-the-art weakly-supervised and unsupervised SOD methods and is even comparable to the fully supervised ones. Code will be released at: https://github.com/czh-3/UADA.

Fatty acid de novo biosynthesis in plastids: Key enzymes and their critical roles for male reproduction and other processes in plants.

Fatty acid de novo biosynthesis in plant plastids is initiated from acetyl-CoA and catalyzed by a series of enzymes, which is required for the vegetative growth, reproductive growth, seed development, stress response, chloroplast development and other biological processes. In this review, we systematically summarized the fatty acid de novo biosynthesis-related genes/enzymes and their critical roles in various plant developmental processes. Based on bioinformatic analysis, we identified fatty acid synthase encoding genes and predicted their potential functions in maize growth and development, especially in anther and pollen development. Finally, we highlighted the potential applications of these fatty acid synthases in male-sterility hybrid breeding, seed oil content improvement, herbicide and abiotic stress resistance, which provides new insights into future molecular crop breeding.

CRISPR/Cas9-based genome editing of 14 lipid metabolic genes reveals a sporopollenin metabolon ZmPKSB-ZmTKPR1-1/-2 required for pollen exine formation in maize.

Lipid biosynthesis and transport are essential for plant male reproduction. Compared with Arabidopsis and rice, relatively fewer maize lipid metabolic genic male-sterility (GMS) genes have been identified, and the sporopollenin metabolon in maize anther remains unknown. Here, we identified two maize GMS genes, ZmTKPR1-1 and ZmTKPR1-2, by CRISPR/Cas9 mutagenesis of 14 lipid metabolic genes with anther stage-specific expression patterns. Among them, tkpr1-1/-2 double mutants displayed complete male sterility with delayed tapetum degradation and abortive pollen. ZmTKPR1-1 and ZmTKPR1-2 encode tetraketide α-pyrone reductases and have catalytic activities in reducing tetraketide α-pyrone produced by ZmPKSB (polyketide synthase B). Several conserved catalytic sites (S128/130, Y164/166 and K168/170 in ZmTKPR1-1/-2) are essential for their enzymatic activities. Both ZmTKPR1-1 and ZmTKPR1-2 are directly activated by ZmMYB84, and their encoded proteins are localized in both the endoplasmic reticulum and nuclei. Based on protein structure prediction, molecular docking, site-directed mutagenesis and biochemical assays, the sporopollenin biosynthetic metabolon ZmPKSB-ZmTKPR1-1/-2 was identified to control pollen exine formation in maize anther. Although ZmTKPR1-1/-2 and ZmPKSB formed a protein complex, their mutants showed different, even opposite, defective phenotypes of anther cuticle and pollen exine. Our findings discover new maize GMS genes that can contribute to male-sterility line-assisted maize breeding and also provide new insights into the metabolon-regulated sporopollenin biosynthesis in maize anther.

Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching.

Weakly supervised semantic segmentation (WSSS) is a challenging yet important research field in vision community. In WSSS, the key problem is to generate high-quality pseudo segmentation masks (PSMs). Existing approaches mainly depend on the discriminative object part to generate PSMs, which would inevitably miss object parts or involve surrounding image background, as the learning process is unaware of the full object structure. In fact, both the discriminative object part and the full object structure are critical for deriving of high-quality PSMs. To fully explore these two information cues, we build a novel end-to-end learning framework, alternate self-dual teaching (ASDT), based on a dual-teacher single-student network architecture. The information interaction among different network branches is formulated in the form of knowledge distillation (KD). Unlike the conventional KD, the knowledge of the two teacher models would inevitably be noisy under weak supervision. Inspired by the Pulse Width (PW) modulation, we introduce a PW wave-like selection signal to alleviate the influence of the imperfect knowledge from either teacher model on the KD process. Comprehensive experiments on the PASCAL VOC 2012 and COCO-Stuff 10K demonstrate the effectiveness of the proposed ASDT framework, and new state-of-the-art results are achieved.

Separating Noisy Samples From Tail Classes for Long-Tailed Image Classification With Label Noise.

Most existing methods that cope with noisy labels usually assume that the classwise data distributions are well balanced. They are difficult to deal with the practical scenarios where training samples have imbalanced distributions, since they are not able to differentiate noisy samples from tail classes' clean samples. This article makes an early effort to tackle the image classification task in which the provided labels are noisy and have a long-tailed distribution. To deal with this problem, we propose a new learning paradigm which can screen out noisy samples by matching between inferences on weak and strong data augmentations. A leave-noise-out regularization (LNOR) is further introduced to eliminate the effect of the recognized noisy samples. Besides, we propose a prediction penalty based on the online classwise confidence levels to avoid the bias toward easy classes which are dominated by head classes. Extensive experiments on five datasets including CIFAR-10, CIFAR-100, MNIST, FashionMNIST, and Clothing1M demonstrate that the proposed method outperforms the existing algorithms for learning with long-tailed distribution and label noise.

Select, Purify, and Exchange: A Multisource Unsupervised Domain Adaptation Method for Building Extraction.

Accurately extracting buildings from aerial images has essential research significance for timely understanding human intervention on the land. The distribution discrepancies between diversified unlabeled remote sensing images (changes in imaging sensor, location, and environment) and labeled historical images significantly degrade the generalization performance of deep learning algorithms. Unsupervised domain adaptation (UDA) algorithms have recently been proposed to eliminate the distribution discrepancies without re-annotating training data for new domains. Nevertheless, due to the limited information provided by a single-source domain, single-source UDA (SSUDA) is not an optimal choice when multitemporal and multiregion remote sensing images are available. We propose a multisource UDA (MSUDA) framework SPENet for building extraction, aiming at selecting, purifying, and exchanging information from multisource domains to better adapt the model to the target domain. Specifically, the framework effectively utilizes richer knowledge by extracting target-relevant information from multiple-source domains, purifying target domain information with low-level features of buildings, and exchanging target domain information in an interactive learning manner. Extensive experiments and ablation studies constructed on 12 city datasets prove the effectiveness of our method against existing state-of-the-art methods, e.g., our method achieves 59.1% intersection over union (IoU) on Austin and Kitsap → Potsdam, which surpasses the target domain supervised method by 2.2% . The code is available at https://github.com/QZangXDU/SPENet.

Multi-Task Learning With Hierarchical Guidance for Locating and Stratifying Submucosal Tumors.

Locating and stratifying the submucosal tumor of the digestive tract from endoscopy ultrasound (EUS) images are of vital significance to the preliminary diagnosis of tumors. However, the above problems are challenging, due to the poor appearance contrast between different layers of the digestive tract wall (DTW) and the narrowness of each layer. Few of existing deep-learning based diagnosis algorithms are devised to tackle this issue. In this article, we build a multi-task framework for simultaneously locating and stratifying the submucosal tumor. And considering the awareness of the DTW is critical to the localization and stratification of the tumor, we integrate the DTW segmentation task into the proposed multi-task framework. Except for sharing a common backbone model, the three tasks are explicitly directed with a hierarchical guidance module, in which the probability map of DTW itself is used to locally enhance the feature representation for tumor localization, and the probability maps of DTW and tumor are jointly employed to locally enhance the feature representation for tumor stratification. Moreover, by means of the dynamic class activation map, probability maps of DTW and tumor are reused to enforce the stratification inference process to pay more attention to DTW and tumor regions, contributing to a reliable and interpretable submucosal tumor stratification model. Additionally, considering the relation with respect to other structures is beneficial for stratifying tumors, we devise a graph reasoning module to replenish non-local relation knowledge for the stratification branch. Experiments on a Stomach-Esophagus and an Intestinal EUS dataset prove that our method achieves very appealing performance on both tumor localization and stratification, significantly outperforming state-of-the-art object detection approaches.

Reliable Mutual Distillation for Medical Image Segmentation Under Imperfect Annotations.

Convolutional neural networks (CNNs) have made enormous progress in medical image segmentation. The learning of CNNs is dependent on a large amount of training data with fine annotations. The workload of data labeling can be significantly relieved via collecting imperfect annotations which only match the underlying ground truths coarsely. However, label noises which are systematically introduced by the annotation protocols, severely hinders the learning of CNN-based segmentation models. Hence, we devise a novel collaborative learning framework in which two segmentation models cooperate to combat label noises in coarse annotations. First, the complementary knowledge of two models is explored by making one model clean training data for the other model. Secondly, to further alleviate the negative impact of label noises and make sufficient usage of the training data, the specific reliable knowledge of each model is distilled into the other model with augmentation-based consistency constraints. A reliability-aware sample selection strategy is incorporated for guaranteeing the quality of the distilled knowledge. Moreover, we employ joint data and model augmentations to expand the usage of reliable knowledge. Extensive experiments on two benchmarks showcase the superiority of our proposed method against existing methods under annotations with different noise levels. For example, our approach can improve existing methods by nearly 3% DSC on the lung lesion segmentation dataset LIDC-IDRI under annotations with 80% noise ratio. Code is available at: https://github.com/Amber-Believe/ReliableMutualDistillation.

Semi-supervised graph convolutional networks for the domain adaptive recognition of thyroid nodules in cross-device ultrasound images.

BACKGROUND: Ultrasound plays a critical role in the early screening and diagnosis of cancers. Although deep neural networks have been widely investigated in the computer-aided diagnosis (CAD) of different medical images, diverse ultrasound devices, and image modalities pose challenges for clinical applications, especially in the recognition of thyroid nodules having various shapes and sizes. More generalized and extensible methods need to be developed for the cross-devices recognition of thyroid nodules. PURPOSE: In this work, a semi-supervised graph convolutional deep learning framework is proposed for the domain adaptative recognition of thyroid nodules across several ultrasound devices. A deep classification network, trained on a source domain with a specific device, can be transferred to recognize thyroid nodules on the target domain with other devices, using only few manual annotated ultrasound images. METHODS: This study presents a semi-supervised graph-convolutional-network-based domain adaptation framework, namely Semi-GCNs-DA. Based on the ResNet backbone, it is extended in three aspects for domain adaptation, that is, graph convolutional networks (GCNs) for the connection construction between source and target domains, semi-supervised GCNs for accurate target domain recognition, and pseudo labels for unlabeled target domains. Data were collected from 1498 patients comprising 12 108 images with or without thyroid nodules under three different ultrasound devices. Accuracy, Sensitivity and Specificity were used for the performance evaluation. RESULTS: The proposed method was validated on six groups of data for a single source domain adaptation task, the mean Accuracy was 0.9719 ± 0.0023, 0.9928 ± 0.0022, 0.9353 ± 0.0105, 0.8727 ± 0.0021, 0.7596 ± 0.0045, 0.8482 ± 0.0092, which achieved better performance in comparison with the state-of-the-art. The proposed method was also validated on three groups of multiple source domain adaptation tasks. In particular, when using X60 and HS50 as the source domain data, and H60 as the target domain, it can achieve the Accuracy of 0.8829 ± 0.0079, Sensitivity of 0.9757 ± 0.0001, and Specificity of 0.7894 ± 0.0164. Ablation experiments also demonstrated the effectiveness of the proposed modules. CONCLUSION: The developed Semi-GCNs-DA framework can effectively recognize thyroid nodules on different ultrasound devices. The developed semi-supervised GCNs can be further extended to the domain adaptation problems for other modalities of medical images.

A Systematic Investigation of Lipid Transfer Proteins Involved in Male Fertility and Other Biological Processes in Maize.

Plant lipid transfer proteins (LTPs) play essential roles in various biological processes, including anther and pollen development, vegetative organ development, seed development and germination, and stress response, but the research progress varies greatly among Arabidopsis, rice and maize. Here, we presented a preliminary introduction and characterization of the whole 65 LTP genes in maize, and performed a phylogenetic tree and gene ontology analysis of the LTP family members in maize. We compared the research progresses of the reported LTP genes involved in male fertility and other biological processes in Arabidopsis and rice, and thus provided some implications for their maize orthologs, which will provide useful clues for the investigation of LTP transporters in maize. We predicted the functions of LTP genes based on bioinformatic analyses of their spatiotemporal expression patterns by using RNA-seq and qRT-PCR assays. Finally, we discussed the advances and challenges in substrate identification of plant LTPs, and presented the future research directions of LTPs in plants. This study provides a basic framework for functional research and the potential application of LTPs in multiple plants, especially for male sterility research and application in maize.

Triphasic regulation of ZmMs13 encoding an ABCG transporter is sequentially required for callose dissolution, pollen exine and anther cuticle formation in maize.

INTRODUCTION: ATP Binding Cassette G (ABCG) transporters are associated with plant male reproduction, while their regulatory mechanisms underlying anther and pollen development remain largely unknown. OBJECTIVES: Identify and characterize a male-sterility gene ZmMs13 encoding an ABCG transporter in modulating anther and pollen development in maize. METHODS: Phenotypic, cytological observations, and histochemistry staining were performed to characterize the ms13-6060 mutant. Map-based cloning and CRISPR/Cas9 gene editing were used to identify ZmMs13 gene. RNA-seq data and qPCR analyses, phylogenetic and microsynteny analyses, transient dual-luciferase reporter and EMSA assays, subcellular localization, and ATPase activity and lipidomic analyses were carried out to determine the regulatory mechanisms of ZmMs13 gene. RESULTS: Maize ms13-6060 mutant displays complete male sterility with delayed callose degradation, premature tapetal programmed cell death (PCD), and defective pollen exine and anther cuticle formation. ZmMs13 encodes a plasm membrane (PM)- and endoplasmic reticulum (ER)-localized half-size ABCG transporter (ZmABCG2a). The allele of ZmMs13 in ms13-6060 mutant has one amino acid (I311) deletion due to a 3-bp deletion in its fourth exon. The I311 and other conserved amino acid K99 are essential for the ATPase and lipid binding activities of ZmMS13. ZmMs13 is specifically expressed in anthers with three peaks at stages S5, S8b, and S10, which are successively regulated by transcription factors ZmbHLH122, ZmMYB84, and ZmMYB33-1/-2 at these three stages. The triphasic regulation of ZmMs13 is sequentially required for callose dissolution, tapetal PCD and pollen exine development, and anther cuticle formation, corresponding to transcription alterations of callose-, ROS-, PCD-, sporopollenin-, and anther cuticle-related genes in ms13-6060 anthers. CONCLUSION: ms13-6060 mutation with one key amino acid (I311) deletion greatly reduces ZmMS13 ATPase and lipid binding activities and displays multiple effects during maize male reproduction. Our findings provide new insights into molecular mechanisms of ABCG transporters controlling anther and pollen development and male fertility in plants.

The ZmMYB84-ZmPKSB regulatory module controls male fertility through modulating anther cuticle-pollen exine trade-off in maize anthers.

Anther cuticle and pollen exine are two crucial lipid layers that ensure normal pollen development and pollen-stigma interaction for successful fertilization and seed production in plants. Their formation processes share certain common pathways of lipid biosynthesis and transport across four anther wall layers. However, molecular mechanism underlying a trade-off of lipid-metabolic products to promote the proper formation of the two lipid layers remains elusive. Here, we identified and characterized a maize male-sterility mutant pksb, which displayed denser anther cuticle but thinner pollen exine as well as delayed tapetal degeneration compared with its wild type. Based on map-based cloning and CRISPR/Cas9 mutagenesis, we found that the causal gene (ZmPKSB) of pksb mutant encoded an endoplasmic reticulum (ER)-localized polyketide synthase (PKS) with catalytic activities to malonyl-CoA and midchain-fatty acyl-CoA to generate triketide and tetraketide α-pyrone. A conserved catalytic triad (C171, H320 and N353) was essential for its enzymatic activity. ZmPKSB was specifically expressed in maize anthers from stages S8b to S9-10 with its peak at S9 and was directly activated by a transcription factor ZmMYB84. Moreover, loss function of ZmMYB84 resulted in denser anther cuticle but thinner pollen exine similar to the pksb mutant. The ZmMYB84-ZmPKSB regulatory module controlled a trade-off between anther cuticle and pollen exine formation by altering expression of a series of genes related to biosynthesis and transport of sporopollenin, cutin and wax. These findings provide new insights into the fine-tuning regulation of lipid-metabolic balance to precisely promote anther cuticle and pollen exine formation in plants.

ATP-Binding Cassette G Transporters and Their Multiple Roles Especially for Male Fertility in Arabidopsis, Rice and Maize.

ATP-binding cassette subfamily G (ABCG) transporters are extensive in plants and play essential roles in various processes influencing plant fitness, but the research progress varies greatly among Arabidopsis, rice and maize. In this review, we present a consolidated nomenclature and characterization of the whole 51 ABCG transporters in maize, perform a phylogenetic analysis and classification of the ABCG subfamily members in maize, and summarize the latest research advances in ABCG transporters for these three plant species. ABCG transporters are involved in diverse processes in Arabidopsis and rice, such as anther and pollen development, vegetative and female organ development, abiotic and biotic stress response, and phytohormone transport, which provide useful clues for the functional investigation of ABCG transporters in maize. Finally, we discuss the current challenges and future perspectives for the identification and mechanism analysis of substrates for plant ABCG transporters. This review provides a basic framework for functional research and the potential application of ABCG transporters in multiple plants, including maize.

Genetic Structure and Molecular Mechanisms Underlying the Formation of Tassel, Anther, and Pollen in the Male Inflorescence of Maize (Zea mays L.).

Maize tassel is the male reproductive organ which is located at the plant's apex; both its morphological structure and fertility have a profound impact on maize grain yield. More than 40 functional genes regulating the complex tassel traits have been cloned up to now. However, the detailed molecular mechanisms underlying the whole process, from male inflorescence meristem initiation to tassel morphogenesis, are seldom discussed. Here, we summarize the male inflorescence developmental genes and construct a molecular regulatory network to further reveal the molecular mechanisms underlying tassel-trait formation in maize. Meanwhile, as one of the most frequently studied quantitative traits, hundreds of quantitative trait loci (QTLs) and thousands of quantitative trait nucleotides (QTNs) related to tassel morphology have been identified so far. To reveal the genetic structure of tassel traits, we constructed a consensus physical map for tassel traits by summarizing the genetic studies conducted over the past 20 years, and identified 97 hotspot intervals (HSIs) that can be repeatedly mapped in different labs, which will be helpful for marker-assisted selection (MAS) in improving maize yield as well as for providing theoretical guidance in the subsequent identification of the functional genes modulating tassel morphology. In addition, maize is one of the most successful crops in utilizing heterosis; mining of the genic male sterility (GMS) genes is crucial in developing biotechnology-based male-sterility (BMS) systems for seed production and hybrid breeding. In maize, more than 30 GMS genes have been isolated and characterized, and at least 15 GMS genes have been promptly validated by CRISPR/Cas9 mutagenesis within the past two years. We thus summarize the maize GMS genes and further update the molecular regulatory networks underlying male fertility in maize. Taken together, the identified HSIs, genes and molecular mechanisms underlying tassel morphological structure and male fertility are useful for guiding the subsequent cloning of functional genes and for molecular design breeding in maize. Finally, the strategies concerning efficient and rapid isolation of genes controlling tassel morphological structure and male fertility and their application in maize molecular breeding are also discussed.

Use of CRISPR/Cas9-Based Gene Editing to Simultaneously Mutate Multiple Homologous Genes Required for Pollen Development and Male Fertility in Maize.

Male sterility represents an important trait for hybrid breeding and seed production in crops. Although the genes required for male fertility have been widely studied and characterized in many plant species, most of them are single genic male-sterility (GMS) genes. To investigate the role of multiple homologous genes in anther and pollen developments of maize, we established the CRISPR/Cas9-based gene editing method to simultaneously mutate the homologs in several putative GMS gene families. By using the integrated strategies of multi-gene editing vectors, maize genetic transformation, mutation-site analysis of T0 and F1 plants, and genotyping and phenotyping of F2 progenies, we further confirmed gene functions of every member in ZmTGA9-1/-2/-3 family, and identified the functions of ZmDFR1, ZmDFR2, ZmACOS5-1, and ZmACOS5-2 in controlling maize male fertility. Single and double homozygous gene mutants of ZmTGA9-1/-2/-3 did not affect anther and pollen development, while triple homozygous gene mutant resulted in complete male sterility. Two single-gene mutants of ZmDFR1/2 displayed partial male sterility, but the double-gene mutant showed complete male sterility. Additionally, only the ZmACOS5-2 single gene was required for anther and pollen development, while ZmACOS5-1 had no effect on male fertility. Our results show that the CRISPR/Cas9 gene editing system is a highly efficient and convenient tool for identifying multiple homologous GMS genes. These findings enrich GMS genes and mutant resources for breeding of maize GMS lines and promote deep understanding of the gene family underlying pollen development and male fertility in maize.

M3Net: A multi-scale multi-view framework for multi-phase pancreas segmentation based on cross-phase non-local attention.

The complementation of arterial and venous phases visual information of CTs can help better distinguish the pancreas from its surrounding structures. However, the exploration of cross-phase contextual information is still under research in computer-aided pancreas segmentation. This paper presents M3Net, a framework that integrates multi-scale multi-view information for multi-phase pancreas segmentation. The core of M3Net is built upon a dual-path network in which individual branches are set up for two phases. Cross-phase interactive connections bridging the two branches are introduced to interleave and integrate dual-phase complementary visual information. Besides, we further devise two types of non-local attention modules to enhance the high-level feature representation across phases. First, we design a location attention module to generate cross-phase reliable feature correlations to suppress the misalignment regions. Second, the depth-wise attention module is used to capture the channel dependencies and then strengthen feature representations. The experiment data consists of 224 internal CTs (106 normal and 118 abnormal) with 1 mm slice thickness, and 66 external CTs (29 normal and 37 abnormal) with 5 mm slice thickness. We achieve new state-of-the-art performance with average DSC of 91.19% on internal data, and promising result with average DSC of 86.34% on external data.

Contralaterally Enhanced Networks for Thoracic Disease Detection.

Identifying and locating diseases in chest X-rays are very challenging, due to the low visual contrast between normal and abnormal regions, and distortions caused by other overlapping tissues. An interesting phenomenon is that there exist many similar structures in the left and right parts of the chest, such as ribs, lung fields and bronchial tubes. This kind of similarities can be used to identify diseases in chest X-rays, according to the experience of broad-certificated radiologists. Aimed at improving the performance of existing detection methods, we propose a deep end-to-end module to exploit the contralateral context information for enhancing feature representations of disease proposals. First of all, under the guidance of the spine line, the spatial transformer network is employed to extract local contralateral patches, which can provide valuable context information for disease proposals. Then, we build up a specific module, based on both additive and subtractive operations, to fuse the features of the disease proposal and the contralateral patch. Our method can be integrated into both fully and weakly supervised disease detection frameworks. It achieves 33.17 AP50 on a carefully annotated private chest X-ray dataset which contains 31,000 images. Experiments on the NIH chest X-ray dataset indicate that our method achieves state-of-the-art performance in weakly-supervised disease localization.

PNEN: Pyramid Non-Local Enhanced Networks.

Existing neural networks proposed for low-level image processing tasks are usually implemented by stacking convolution layers with limited kernel size. Every convolution layer merely involves in context information from a small local neighborhood. More contextual features can be explored as more convolution layers are adopted. However it is difficult and costly to take full advantage of long-range dependencies. We propose a novel non-local module, Pyramid Non-local Block, to build up connection between every pixel and all remain pixels. The proposed module is capable of efficiently exploiting pairwise dependencies between different scales of low-level structures. The target is fulfilled through first learning a query feature map with full resolution and a pyramid of reference feature maps with downscaled resolutions. Then correlations with multi-scale reference features are exploited for enhancing pixel-level feature representation. The calculation procedure is economical considering memory consumption and computational cost. Based on the proposed module, we devise a Pyramid Non-local Enhanced Networks for edge-preserving image smoothing which achieves state-of-the-art performance in imitating three classical image smoothing algorithms. Additionally, the pyramid non-local block can be directly incorporated into convolution neural networks for other image restoration tasks. We integrate it into two existing methods for image denoising and single image super-resolution, achieving consistently improved performance.

Self-Enhanced Convolutional Network for Facial Video Hallucination.

As a domain-specific super-resolution problem, facial image hallucination has enjoyed a series of breakthroughs thanks to the advances of deep convolutional neural networks. However, the direct migration of existing methods to video is still difficult to achieve good performance due to its lack of alignment and consistency modelling in temporal domain. Taking advantage of high inter-frame dependency in videos, we propose a selfenhanced convolutional network for facial video hallucination. It is implemented by making full usage of preceding super-resolved frames and a temporal window of adjacent low-resolution frames. Specifically, the algorithm first obtains the initial high-resolution inference of each frame by taking into consideration a sequence of consecutive low-resolution inputs through temporal consistency modelling. It further recurrently exploits the reconstructed results and intermediate features of a sequence of preceding frames to improve the initial super-resolution of the current frame by modelling the coherence of structural facial features across frames. Quantitative and qualitative evaluations demonstrate the superiority of the proposed algorithm against state-of-theart methods. Moreover, our algorithm also achieves excellent performance in the task of general video super-resolution in a single-shot setting.

Identification of S23 causing both interspecific hybrid male sterility and environment-conditioned male sterility in rice.

BACKGROUND: Oryza glumaepatula represents an important resource of genetic diversity that can be used to improve rice production. However, hybrid sterility severely restricts gene flow between Oryza species, and hinders the utilization of distant heterosis in hybrid rice breeding. RESULTS: In order to fully exploit the beneficial genes of O. glumaepatula and facilitate the conservation of these gene resources, a set of chromosome single-segment substitution lines (SSSLs) was developed using an indica variety HJX74 as the recurrent parent and an accession of O. glumaepatula as the donor parent. During the process of SSSLs development, S23, a locus conferring hybrid male sterility between O. sativa and O. glumaepatula, was identified and fine mapped to 11.54 kb and 7.08 kb genomic region in O. sativa and O. glumaepatula, respectively, encoding three and two candidate ORFs, respectively. qRT-PCR and sequence analysis excluded one common ORF as the candidate gene. In addition, hybrid male sterility caused by S23 was environment-sensitive, and could be observed only in natural short-day (NSD). CONCLUSION: Identification and candidate genes analysis of S23 in this study provides a valuable example to study the crosstalk between interspecific F1 hybrid male sterility and environment-conditioned male sterility in rice, facilitates reserving and utilizing favorable genes or alleles of wild Oryza species, and allows for a more efficient exploitation of distant heterosis in hybrid rice breeding.