Image-to-Character-to-Word Transformers for Accurate Scene Text Recognition.

Leveraging the advances of natural language processing, most recent scene text recognizers adopt an encoder-decoder architecture where text images are first converted to representative features and then a sequence of characters via 'sequential decoding'. However, scene text images suffer from rich noises of different sources such as complex background and geometric distortions which often confuse the decoder and lead to incorrect alignment of visual features at noisy decoding time steps. This paper presents I2C2W, a novel scene text recognition technique that is tolerant to geometric and photometric degradation by decomposing scene text recognition into two inter-connected tasks. The first task focuses on image-to-character (I2C) mapping which detects a set of character candidates from images based on different alignments of visual features in an non-sequential way. The second task tackles character-to-word (C2W) mapping which recognizes scene text by decoding words from the detected character candidates. The direct learning from character semantics (instead of noisy image features) corrects falsely detected character candidates effectively which improves the final text recognition accuracy greatly. Extensive experiments over nine public datasets show that the proposed I2C2W outperforms the state-of-the-art by large margins for challenging scene text datasets with various curvature and perspective distortions. It also achieves very competitive recognition performance over multiple normal scene text datasets.

Conditional Hyper-Network for Blind Super-Resolution With Multiple Degradations.

Although the single-image super-resolution (SISR) methods have achieved great success on the single degradation, they still suffer performance drop with multiple degrading effects in real scenarios. Recently, some blind and non-blind models for multiple degradations have been explored. However, these methods usually degrade significantly for distribution shifts between the training and test data. Towards this end, we propose a novel conditional hyper-network framework for super-resolution with multiple degradations (named CMDSR), which helps the SR framework learn how to adapt to changes in the degradation distribution of input. We extract degradation prior at the task-level with the proposed ConditionNet, which will be used to adapt the parameters of the basic SR network (BaseNet). Specifically, the ConditionNet of our framework first learns the degradation prior from a support set, which is composed of a series of degraded image patches from the same task. Then the adaptive BaseNet rapidly shifts its parameters according to the conditional features. Moreover, in order to better extract degradation prior, we propose a task contrastive loss to shorten the inner-task distance and enlarge the cross-task distance between task-level features. Without predefining degradation maps, our blind framework can conduct one single parameter update to yield considerable improvement in SR results. Extensive experiments demonstrate the effectiveness of CMDSR over various blind, and even several non-blind methods. The flexible BaseNet structure also reveals that CMDSR can be a general framework for a large series of SISR models. Our code is available at https://github.com/guanghaoyin/CMDSR.

Root microbiome changes associated with cadmium exposure and/or overexpression of a transgene that reduces Cd content in rice.

Cadmium (Cd) is a toxic heavy metal that can accumulate in crop plants. We reported previously the engineering of a low cadmium-accumulating line (2B) of rice through overexpression of a truncated OsO3L2 gene. As expression of this transgene was highest in plant roots, amplicon and metatranscriptome sequencing were used to investigate the possibility that its expression affects root associated microbes. Based on amplicon sequencing of bacterial 16S rRNA, but less so from fungal ITS, the OTUs (operational taxonomic units) showed less diversity in soil tightly (rhizoplane) than loosely (rhizosphere) associated with plant roots. Significantly changed OTUs caused by the low-Cd accumulating plant 2B, Cd treatment or both were found, and 10 of the 13 OTUs (77%) that were enriched in Cd treated 2B samples over the wild type counterpart have been previously described as involved in tolerance to Cd or other heavy metals. Metatranscriptome sequencing of rhizosphere microbiome found that bacteria accounted for 70-75% of the microbial RNA. Photosynthesis-antenna proteins and nitrogen metabolism pathways were most active in soil microbes treated with Cd and grown with plant 2B. Correspondingly, the relative abundance of Cyanobacteria was enriched to < 1% of Cd treated rhizosphere bacteria, yet accounted for up to 13% of Cd treated 2B rhizospheric transcripts. These enriched microbes by transgene and Cd are worthy candidates for future application on reducing crop uptake of Cd.

Birds of a Feather Flock Together: Category-Divergence Guidance for Domain Adaptive Segmentation.

Unsupervised domain adaptation (UDA) aims to enhance the generalization capability of a certain model from a source domain to a target domain. Present UDA models focus on alleviating the domain shift by minimizing the feature discrepancy between the source domain and the target domain but usually ignore the class confusion problem. In this work, we propose an Inter-class Separation and Intra-class Aggregation (ISIA) mechanism. It encourages the cross-domain representative consistency between the same categories and differentiation among diverse categories. In this way, the features belonging to the same categories are aligned together and the confusable categories are separated. By measuring the align complexity of each category, we design an Adaptive-weighted Instance Matching (AIM) strategy to further optimize the instance-level adaptation. Based on our proposed methods, we also raise a hierarchical unsupervised domain adaptation framework for cross-domain semantic segmentation task. Through performing the image-level, feature-level, category-level and instance-level alignment, our method achieves a stronger generalization performance of the model from the source domain to the target domain. In two typical cross-domain semantic segmentation tasks, i.e., GTA 5→ Cityscapes and SYNTHIA → Cityscapes, our method achieves the state-of-the-art segmentation accuracy. We also build two cross-domain semantic segmentation datasets based on the publicly available data, i.e., remote sensing building segmentation and road segmentation, for domain adaptive segmentation. Our code, models and datasets are available at https://github.com/HibiscusYB/BAFFT.

Fine-Grained Human-Centric Tracklet Segmentation with Single Frame Supervision.

In this paper, we target at the Fine-grAined human-Centric Tracklet Segmentation (FACTS) problem, where 12 human parts, e.g., face, pants, left-leg, are segmented. To reduce the heavy and tedious labeling efforts, FACTS requires only one labeled frame per video during training. The small size of human parts and the labeling scarcity makes FACTS very challenging. Considering adjacent frames of videos are continuous and human usually do not change clothes in a short time, we explicitly consider the pixel-level and frame-level context in the proposed Temporal Context segmentation Network (TCNet). On the one hand, optical flow is on-line calculated to propagate the pixel-level segmentation results to neighboring frames. On the other hand, frame-level classification likelihood vectors are also propagated to nearby frames. By fully exploiting the pixel-level and frame-level context, TCNet indirectly uses the large amount of unlabeled frames during training and produces smooth segmentation results during inference. Experimental results on four video datasets show the superiority of TCNet over the state-of-the-arts. The newly annotated datasets can be downloaded via http://liusi-group.com/projects/FACTS for the further studies.

Arabidopsis OXIDATIVE STRESS 3 enhances stress tolerance in Schizosaccharomyces pombe by promoting histone subunit replacement that upregulates drug-resistant genes.

Histone replacement in chromatin-remodeling plays an important role in eukaryotic gene expression. New histone variants replacing their canonical counterparts often lead to a change in transcription, including responses to stresses caused by temperature, drought, salinity, and heavy metals. In this study, we describe a chromatin-remodeling process triggered by eviction of Rad3/Tel1-phosphorylated H2Aα, in which a heterologous plant protein AtOXS3 can subsequently bind fission yeast HA2.Z and Swc2, a component of the SWR1 complex, to facilitate replacement of H2Aα with H2A.Z. The histone replacement increases occupancy of the oxidative stress-responsive transcription factor Pap1 at the promoters of at least three drug-resistant genes, which enhances their transcription and hence primes the cell for higher stress tolerance.

Arabidopsis OXS3 family proteins repress ABA signaling through interactions with AFP1 in the regulation of ABI4 expression.

Abscisic acid (ABA) and the AP2/ERF (APETALA2/ETHYLENE-RESPONSIVE FACTOR)-type transcription factor called ABA INSENSITIVE 4 (ABI4) play pivotal roles in plant growth responses to environmental stress. An analysis of seedling development in Arabidopsis ABA hypersensitive mutants suggested that OXS3 (OXIDATIVE STRESS 3), OXS3b, O3L3 (OXS3 LIKE 3), O3L4, and O3L6 were negative regulators of ABI4 expression. We therefore characterized the roles of the OXS3 family members in ABA signaling. All the above five OXS3 proteins were found to interact with AFP1 (ABI FIVE BINDING PROTEIN 1) in yeast two hybrid assays. Seven OXS3 family members including two other members O3L1 and O3L5 were found to interact with histone H2A.X, although OXS3b, O3L3, and O3L5 showed weaker interactions. ChIP-qPCR analysis showed that the absence of some of these OXS3 family proteins was associated with increased occupancy of histone γ-H2A.X at the ABI4 promoter, which also corresponded with de-repression of ABI4 expression. Repression of ABI4 expression, however, required both AFP1 and OXS3, OXS3b or O3L6. We conclude that in the absence of stress, OXS3 family proteins regulate γ-H2A.X deposition at the ABI4 promoter and that together with AFP1, OXS3 family proteins function to prevent ABA-induced growth arrest by co-repressing ABI4 through decreased promoter occupancy of histone γ-H2A.X.

SnRK1 regulates chromatin-associated OXS3 family proteins localization through phosphorylation in Arabidopsis thaliana.

In plants, SNF1-related protein kinase 1 (SnRK1) senses nutrient and energy status and transduces this information into appropriate responses. Oxidative Stress 3 (OXS3) and family members share a highly conserved putative N-acetyltransferase catalytic domain (ACD). Here, we describe that the ACD contains two candidate SnRK1 recognition motifs and that SnRK1 can interact with most of the OXS3 family proteins. In vitro, SnRK1.1 can phosphorylate OXS3, OXS3b and O3L4, and in vivo promote the translocation of OXS3, OXS3b and O3L6 from the nucleus to the cytoplasm. Phosphorylation sites within the OXS3 ACD affect OXS3 cytoplasmic accumulation, as well as their interactions with SnRK1.1. This suggests that signal transduction from SnRK1 to OXS3 family proteins, and that SnRK1 can control their activities through phosphorylation-induced nuclear exclusion.

Category-Aware Spatial Constraint for Weakly Supervised Detection.

Weakly supervised object detection has attracted increasing research attention recently. To this end, most existing schemes rely on scoring category-independent region proposals, which is formulated as a multiple instance learning problem. During this process, the proposal scores are aggregated and supervised by only image-level labels, which often fails to locate object boundaries precisely. In this paper, we break through such a restriction by taking a deeper look into the score aggregation stage and propose a Category-aware Spatial Constraint (CSC) scheme for proposals, which is integrated into weakly supervised object detection in an end-to-end learning manner. In particular, we incorporate the global shape information of objects as an unsupervised constraint, which is inferred from build-in foreground-and-background cues, termed Category-specific Pixel Gradient (CPG) maps. Specifically, each region proposal is weighted according to how well it covers the estimated shape of objects. For each category, a multi-center regularization is further introduced to penalize the violations between centers cluster and high-score proposals in a given image. Extensive experiments are done on the most widely-used benchmark Pascal VOC and COCO, which shows that our approach significantly improves weakly supervised object detection without adding new learnable parameters to the existing models nor changing the structures of CNNs.

Engineering low-cadmium rice through stress-inducible expression of OXS3-family member genes.

Cadmium (Cd) as a carcinogen poses a great threat to food security and public health through plant-derived foods such as rice, the staple for nearly half of the world's population. We have previously reported that overexpression of truncated gene fragments derived from the rice genes OsO3L2 and OsO3L3 could reduce Cd accumulation in transgenic rice. However, we did not test the full length genes due to prior work in Arabidopsis where overexpression of these genes caused seedling lethality. Here, we report on limiting the overexpression of OsO3L2 and OsO3L3 through the use of the stress- inducible promoter RD29B. However, despite generating 625 putative transformants, only 7 lines survived as T1 seedlings and only 1 line of each overexpressed OsO3L2 or OsO3L3-produced T2 progeny. The T2 homozygotes from these 2 lines showed the same effect of reducing accumulation of Cd in root and shoot as well as in T3 grain. As importantly, the concentrations of essential metals copper (Cu), iron (Fe), manganese (Mn) and zinc (Zn) were unaffected. Analysis of the expression profile suggested that low Cd accumulation may be due to high expression of OsO3L2 and OsO3L3 in the root tip region. Cellular localization of OsO3L2 and OsO3L3 indicate that they are histone H2A interacting nuclear proteins in vascular cells and especially in the root tip region. It is possible that interaction with histone H2A modifies chromatin to regulate downstream gene expression.

Weakly Supervised Object Detection via Object-Specific Pixel Gradient.

Most existing object detection algorithms are trained based upon a set of fully annotated object regions or bounding boxes, which are typically labor-intensive. On the contrary, nowadays there is a significant amount of image-level annotations cheaply available on the Internet. It is hence a natural thought to explore such "weak" supervision to benefit the training of object detectors. In this paper, we propose a novel scheme to perform weakly supervised object localization, termed object-specific pixel gradient (OPG). The OPG is trained by using image-level annotations alone, which performs in an iterative manner to localize potential objects in a given image robustly and efficiently. In particular, we first extract an OPG map to reveal the contributions of individual pixels to a given object category, upon which an iterative mining scheme is further introduced to extract instances or components of this object. Moreover, a novel average and max pooling layer is introduced to improve the localization accuracy. In the task of weakly supervised object localization, the OPG achieves a state-of-the-art 44.5% top-5 error on ILSVRC 2013, which outperforms competing methods, including Oquab et al. and region-based convolutional neural networks on the Pascal VOC 2012, with gains of 2.6% and 2.3%, respectively. In the task of object detection, OPG achieves a comparable performance of 27.0% mean average precision on Pascal VOC 2007. In all experiments, the OPG only adopts the off-the-shelf pretrained CNN model, without using any object proposals. Therefore, it also significantly improves the detection speed, i.e., achieving three times faster compared with the state-of-the-art method.