We compute topological correlators in Landau-Ginzburg models on a Riemann surface with arbitrary number of handles and boundaries. The boundaries may correspond to arbitrary topological D-branes of type B. We also allow arbitrary operator insertions on the boundary and in the bulk. The answer is given by an explicit formula which can be regarded as an open-string generalization of C. Vafa's formula for closed-string topological correlators. We discuss how to extend our results to the case of Landau-Ginzburg orbifolds.

We study the topological sector of N = 2 sigma-models with Hflux. It has been known for a long time that the target-space geometry of these theories is not Kähler and can be described in terms of a pair of complex structures, which do not commute, in general, and are parallel with respect to two different connections with torsion. Recently an alternative description of this geometry was found, which involves a pair of commuting twisted generalized complex structures on the target space. In this paper we define and study the analogues of A and B-models for N = 2 sigma-models with H-flux and show that the results are naturally expressed in the language of twisted generalized complex geometry. For example, the space of topological observables is given by the cohomology of a Lie algebroid associated to one of the two twisted generalized complex structures. We determine the topological scalar product, which endows the algebra of observables with the structure of a Frobenius algebra. We also discuss mirror symmetry for twisted generalized Calabi-Yau manifolds. CALT-68-25141 More precisely, the B-model makes sense on the quantum level if and only if M is a Calabi-Yau manifold. For the A-model, the Calabi-Yau condition is unnecessary.

Text line segmentation in freestyle handwritten documents remains an open document analysis problem. Curvilinear text lines and small gaps between neighboring text lines present a challenge to algorithms developed for machine printed or hand-printed documents. In this paper, we propose a novel approach based on density estimation and a state-of-the-art image segmentation technique, the level set method. From an input document image, we estimate a probability map, where each element represents the probability that the underlying pixel belongs to a text line. The level set method is then exploited to determine the boundary of neighboring text lines by evolving an initial estimate. Unlike most connected component based methods [1,2], the proposed algorithm does not use any script-specific knowledge. Extensive quantitative experiments on freestyle handwritten documents with diverse scripts, such as Arabic, Chinese, Korean, and Hindi, demonstrate that our algorithm consistently outperforms previous methods [3,1,2]. Further experiments show the proposed algorithm is robust to scale change, rotation, and noise. Keywords: Handwritten Text Line Segmentation, Document Image Analysis, Density Estimation, Level Set MethodsThe support of this research by the Department of Defense under contract MDA-9040-2C-0406 is gratefully acknowledged. 1 Report Documentation PageForm Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

Steel strip surface defects recognition is very important to steel strip production and quality control, which needs further improvement. In this paper, an end‐to‐end surface defects recognition system is proposed for steel strip surface inspection. This system is based on the symmetric surround saliency map for surface defects detection and deep convolutional neural networks (CNNs) which directly use the defect image as input and defect category as output for seven classes of steel strip defects classification. The CNNs are trained purely on raw defect images and learned defect features from the training of network, which avoiding the separation between feature extraction and image classification, so that forms an end‐to‐end defects recognition pipeline. To further illustrate the superiority of the defect recognition methods with CNNs, an authoritative and standard steel strip surface defect dataset − NEU is also used to evaluate the defect recognition effect using CNNs. Experimental results demonstrate that the proposed methods perform well in steel strip surface defect detection of different types and achieve a high recognition rate for defect images. In addition, a series of data augmentation methods are discussed to analyze its effect on avoiding over‐fitting for defects recognition.

BackgroundOral squamous cell carcinoma (OSCC) is becoming more common across the globe. The prognosis of OSCC is largely dependent on the early detection. But the routine oral cavity examination may delay the diagnosis because the early oral malignant lesions may be clinically indistinguishable from benign or inflammatory diseases. In this study, the new diagnostic method is developed by using the surface enhanced Raman spectroscopy (SERS) to detect the serum samples from the cancer patients.MethodThe blood serum samples were collected from the OSCC patients, MEC patients and the volunteers without OSCC or MEC. Gold nanoparticles(NPs) were then mixed in the serum samples to obtain the high quality SERS spectra. There were totally 135 spectra of OSCC, 90 spectra of mucoepidermoid carcinoma (MEC) and 145 spectra of normal control group, which were captured by SERS successfully. Compared with the normal control group, the Raman spectral differences exhibited in the spectra of OSCC and MEC groups, which were assigned to the nucleic acids, proteins and lipids. Based on these spectral differences and features, the algorithms of principal component analysis(PCA) and linear discriminant analysis (LDA) were employed to analyze and classify the Raman spectra of different groups.ResultsCompared with the normal groups, the major increased peaks in the OSCC and MEC groups were assigned to the molecular structures of the nucleic acids and proteins. And these different major peaks between the OSCC and MEC groups were assigned to the special molecular structures of the carotenoids and lipids. The PCA-LDA results demonstrated that OSCC could be discriminated successfully from the normal control groups with a sensitivity of 80.7% and a specificity of 84.1%. The process of the cross validation proved the results analyzed by PCA-LDA were reliable.ConclusionThe gold NPs were appropriate substances to capture the high-quality SERS spectra of the OSCC, MEC and normal serum samples. The results of this study confirm that SERS combined PCA-LDA had a giant capability to detect and diagnosis OSCC through the serum sample successfully.

The development of in-memory computing has opened up possibilities to build next-generation non-von-Neumann computing architecture. Implementation of logic functions within the memristors can significantly improve the energy efficiency and alleviate the bandwidth congestion issue. In this work, the demonstration of arithmetic logic unit functions is presented in a memristive crossbar with implemented non-volatile Boolean logic and arithmetic computing. For logic implementation, a standard operating voltage mode is proposed for executing reconfigurable stateful IMP, destructive OR, NOR, and non-destructive OR logic on both the word and bit lines. No additional voltages are needed beyond "V P " and its negative component. With these basic logic functions, other Boolean functions are constructed within five devices in at most five steps. For arithmetic computing, the fundamental functions including an n-bit full adder with high parallelism as well as efficient increment, decrement, and shift operations are demonstrated. Other arithmetic blocks, such as subtraction, multiplication, and division are further designed. This work provides solid evidence that memristors can be used as the building block for in-memory computing, targeting various low-power edge computing applications.

No abstract

SF, the most widely used electrical-equipment-insulation gas, has serious greenhouse effects. CFO has attracted much attention as an alternative gas in recent two years, but the environmental impact of its decomposition products is unclear. In this work, the decomposition characteristics of CFO were studied based on gas chromatography-mass spectrometry and density functional theory. We found that the amount of decomposition products of CFO, namely, CF, CF, CF, CF, CF, and CF, increased with increased number of discharges. Under a high-energy electric field, the C-C bond of CFO between carbonyl carbon and α-carbon atoms was most likely to break and generate CFCO•, CF• or CFCO•, CF• free radicals. CF•, and CF• free radicals produced by the breakage more easily recombined to form small molecular products. By analyzing the ionization parameters, toxicity, and environmental effects of CFO and its decomposition products, we found that CFO gas mixtures exhibit great decomposition and environmental characteristics with low toxicity, with great potential to replace SF.