Liver metastasis, the leading cause of colorectal cancer mortality, exhibits a highly heterogeneous and suppressive immune microenvironment. Here, we sequenced 97 matched samples by using single-cell ...RNA sequencing and spatial transcriptomics. Strikingly, the metastatic microenvironment underwent remarkable spatial reprogramming of immunosuppressive cells such as
M2-like macrophages. We further developed scMetabolism, a computational pipeline for quantifying single-cell metabolism, and observed that those macrophages harbored enhanced metabolic activity. Interestingly, neoadjuvant chemotherapy could block this status and restore the antitumor immune balance in responsive patients, whereas the nonresponsive patients deteriorated into a more suppressive one. Our work described the immune evolution of metastasis and uncovered the black box of how tumors respond to neoadjuvant chemotherapy. SIGNIFICANCE: We present a single-cell and spatial atlas of colorectal liver metastasis and found the highly metabolically activated
M2-like macrophages in metastatic sites. Efficient neoadjuvant chemotherapy can slow down such metabolic activation, raising the possibility to target metabolism pathways in metastasis.
.
Scene text recognition has been a hot research topic in computer vision due to its various applications. The state of the art is the attention-based encoder-decoder framework that learns the mapping ...between input images and output sequences in a purely data-driven way. However, we observe that existing attention-based methods perform poorly on complicated and/or low-quality images. One major reason is that existing methods cannot get accurate alignments between feature areas and targets for such images. We call this phenomenon "attention drift". To tackle this problem, in this paper we propose the FAN (the abbreviation of Focusing Attention Network) method that employs a focusing attention mechanism to automatically draw back the drifted attention. FAN consists of two major components: an attention network (AN) that is responsible for recognizing character targets as in the existing methods, and a focusing network (FN) that is responsible for adjusting attention by evaluating whether AN pays attention properly on the target areas in the images. Furthermore, different from the existing methods, we adopt a ResNet-based network to enrich deep representations of scene text images. Extensive experiments on various benchmarks, including the IIIT5k, SVT and ICDAR datasets, show that the FAN method substantially outperforms the existing methods.
Recognizing text from natural images is a hot research topic in computer vision due to its various applications. Despite the enduring research of several decades on optical character recognition ...(OCR), recognizing texts from natural images is still a challenging task. This is because scene texts are often in irregular (e.g. curved, arbitrarily-oriented or seriously distorted) arrangements, which have not yet been well addressed in the literature. Existing methods on text recognition mainly work with regular (horizontal and frontal) texts and cannot be trivially generalized to handle irregular texts. In this paper, we develop the arbitrary orientation network (AON) to directly capture the deep features of irregular texts, which are combined into an attention-based decoder to generate character sequence. The whole network can be trained end-to-end by using only images and word-level annotations. Extensive experiments on various benchmarks, including the CUTE80, SVT-Perspective, IIIT5k, SVT and ICDAR datasets, show that the proposed AON-based method achieves the-state-of-the-art performance in irregular datasets, and is comparable to major existing methods in regular datasets.
The adsorption capacity of oyster shell powders (SPs) and the adsorption mechanism of heavy metal ions (HMs; i.e., cadmium ions Cd
2+
and lead ions Pb
2+
) on SPs are discussed by means of adsorption ...kinetics tests, adsorption-desorption tests, scanning electron microscopy and Fourier transform infrared spectroscopy. The influences of seepage velocity, heavy metal types, and SP addition amount/concentration on the adsorption effect of SPs in the treatment of HMs in laterite as well as quartz sand were analyzed. Studies have shown that i) the adsorption of HMs on SPs can be divided into three stages, i.e., the surface adsorption stage, the internal pore diffusion stage, and the equilibrium stage; ii) with the increase in seepage velocity, the effluent concentration of HMs will slightly increase, and the residual amounts at each section of the column generally decrease rapidly with the increase in migration distance; iii) the increase in the concentration of SP solution provides more adsorption points for the adsorption of HMs, and finally, the amount of HMs desorbed from quartz sand is reduced, which also reduces the concentration of HMs in the effluent. Overall, SPs possess high purification efficiency for the HMs of contaminated soils.
Edit Probability for Scene Text Recognition Bai, Fan; Cheng, Zhanzhan; Niu, Yi ...
2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,
06/2018
Conference Proceeding
We consider the scene text recognition problem under the attention-based encoder-decoder framework, which is the state of the art. The existing methods usually employ a frame-wise maximal likelihood ...loss to optimize the models. When we train the model, the misalignment between the ground truth strings and the attention's output sequences of probability distribution, which is caused by missing or superfluous characters, will confuse and mislead the training process, and consequently make the training costly and degrade the recognition accuracy. To handle this problem, we propose a novel method called edit probability (EP) for scene text recognition. EP tries to effectively estimate the probability of generating a string from the output sequence of probability distribution conditioned on the input image, while considering the possible occurrences of missing/superfluous characters. The advantage lies in that the training process can focus on the missing, superfluous and unrecognized characters, and thus the impact of the misalignment problem can be alleviated or even overcome. We conduct extensive experiments on standard benchmarks, including the IIIT-5K, Street View Text and ICDAR datasets. Experimental results show that the EP can substantially boost scene text recognition performance.
This paper investigates the resilient fault-tolerant model-free adaptive platooning security control (FT-MFAPSC) issue for the vehicular platooning systems (VPSs) subject to sensor faults and ...aperiodic denial-of-service (DoS) attacks. Firstly, an equivalent linear data model can be obtained from the nonlinear VPSs based on the partial form dynamic linearization (PFDL) technique. Then, the fault-tolerant control framework is developed with consideration of the sensor faults. The gradient descent method-based neural network is adopted in the control framework for the fault approximation. Thirdly, an attack compensation mechanism is designed for the PFDL-based controller. Aiming at the VPSs against aperiodic DoS attacks, a novel resilient FT-MFAPSC algorithm with the compensation mechanism is proposed. And the control objectives of the VPSs can be accomplished. Finally, by the simulation example with comparisons, the effectiveness of the developed algorithm can be illustrated.
The demand for vehicular mobile data services has increased exponentially, which necessitates alternative data pipes for vehicular users other than the cellular network and dedicated short-range ...communication. In this paper, we study the performance of underlaid vehicular device-to-device (V-D2D) communications, where the cellular uplink resources are reused by V-D2D communications, considering the characteristics of the vehicular network. Specifically, we model the considered urban area by a grid-like street layout, with nonhomogeneous distribution of vehicle density. We then propose to employ a joint power control and mode selection scheme for the V-D2D communications. In the scheme, we use channel inversion to control the transmit power, in order to determine transmit power based on path loss rather than instantaneous channel state information (CSI), and avoid severe interference due to excessively large transmit power; the transmission mode is selected based on the biased channel quality, where D2D mode is chosen when the biased D2D link quality is not worse than the cellular uplink quality. Under the proposed scheme, two performance metrics of V-D2D underlaid cellular networks, i.e., signal-to-interference-plus-noise outage probability and link/network throughput, are theoretically analyzed. Simulation results validate our analysis and show the impacts of design parameters on the network performance.
Real-time path planning can efficiently relieve traffic congestion in urban scenarios. However, how to design an efficient path-planning algorithm to achieve a globally optimal vehicle-traffic ...control still remains a challenging problem, particularly when we take drivers' individual preferences into consideration. In this paper, we first establish a hybrid intelligent transportation system (ITS), i.e., a hybrid-VANET-enhanced ITS, which utilizes both vehicular ad hoc networks (VANETs) and cellular systems of the public transportation system to enable real-time communications among vehicles, roadside units (RSUs), and a vehicle-traffic server in an efficient way. Then, we propose a real-time path-planning algorithm, which not only improves the overall spatial utilization of a road network but reduces average vehicle travel cost for avoiding vehicles from getting stuck in congestion as well. A stochastic Lyapunov optimization technique is exploited to address the globally optimal path-planning problem. Finally, the transmission delay of the hybrid-VANET-enhanced ITS is evaluated in VISSIM to show the timeliness of the proposed communication framework. Moreover, system-level simulations conducted in Java demonstrate that the proposed path-planning algorithm outperforms the traditional distributed path planning in terms of balancing the spatial utilization and drivers' travel cost.
Knowledge of somatic mutation accumulation in normal cells, which is essential for understanding cancer development and evolution, remains largely lacking. In this study, we investigated somatic ...clonal events in morphologically normal human urothelium (MNU; epithelium lining the bladder and ureter) and identified macroscopic clonal expansions. Aristolochic acid (AA), a natural herb-derived compound, was a major mutagenic driving factor in MNU. AA drastically accelerates mutation accumulation and enhances clonal expansion. Mutations in MNU were widely observed in chromatin remodeling genes such as
and
but rarely in
,
, and
mutations were found to be common in urothelial cells, regardless of whether the cells experience exogenous mutagen exposure. Copy number alterations were rare and largely confined to small-scale regions, along with copy-neutral loss of heterozygosity. Single AA-associated clones in MNU expanded to a scale of several square centimeters in size.
Background & Aims T-helper (Th)17 cells that secrete interleukin (IL)-22 have immunomodulatory and protective properties in the liver and other tissues. IL-22 induces expression of proinflammatory ...genes but is also mitogenic and antiapoptotic in hepatocytes. Therefore, it could have multiple functions in the immune response to hepatitis B virus (HBV). Methods We examined the role of IL-22 in regulating liver inflammation in HBV transgenic mice and measured levels of IL-22 in HBV-infected patients. Results In HBV transgenic mice, injection of a single dose of IL-22 increased hepatic expression of proinflammatory genes but did not directly inhibit virus replication. When splenocytes from HBV-immunized mice were transferred into HBV transgenic mice, the severity of the subsequent liver damage was ameliorated by neutralization of IL-22. In this model, IL-22 depletion did not affect interferon gamma–mediated noncytopathic inhibition of virus replication initiated by HBV-specific cytotoxic T cells, but it significantly inhibited recruitment of antigen-nonspecific inflammatory cells into the liver. In patients with acute HBV infections, the percentage of Th17 cells in peripheral blood and concentration of IL-22 in serum were significantly increased. Conclusions IL-22 appears to be an important mediator of the inflammatory response following recognition of HBV by T cells in the liver. These findings might be relevant to the development of cytokine-based therapies for patients with HBV infection.