In many microarray studies, classifiers have been constructed based on gene signatures to predict clinical outcomes for various cancer sufferers. However, originating from different studies often suffer poor robustness when used in the classification of data sets independent which they were generated from. this paper, we present an unsupervised feature learning framework by integrating a principal component analysis algorithm and autoencoder neural network identify characteristics expression...

This paper presents 6D vision transformer (6D-ViT), a transformer-based instance representation learning network suitable for highly accurate category-level object pose estimation based on RGB-D images. Specifically, novel two-stream encoder-decoder framework is dedicated to exploring complex and powerful representations from RGB images, point clouds, categorical shape priors. The whole consists of two main branches, named Pixelformer Pointformer. contains pyramid encoder with an...

Objective To investigate death rates in schizophrenia and related psychoses. Design Data from two epidemiologically complete cohorts of patients presenting for the first time to mental health services North Wales whom there are at least 1, up 10-year follow-up data have been used calculate survival standardised Setting The Asylum Denbigh (archived patient case notes) West District General Hospital psychiatric unit. Population Cohort 1: notes). Of 3168 admitted 1875–1924, 1074 had a...

The geometric and semantic information of 3D point clouds significantly influence the analysis cloud structures. However, learning based on deep is challenging due to naturally unordered data structure. In this work, we strive impart machines with knowledge object shapes, thereby enabling them infer high-level from model. Inspired by vector locally aggregated descriptors, propose indirectly describing associating each point's low-level descriptor a few visual words. Based approach, design an...

As China’s railways continue to expand into the Qinghai–Tibet Plateau, number of deep-buried long tunnels is increasing. Tunnel-damaging geothermal disasters have become a common problem in underground engineering. Predicting potential disaster areas along Yunnan–Tibet railway project conducive its planning and construction realization United Nations Sustainable Development Goals (SDGs)—specifically, industry, innovation infrastructure goal (SDG 9). In this paper, was study area. Landsat-8...

The geothermal resources in the southwest section of Mid-Spine Belt Beautiful China are abundant, but quantitative prediction and evaluation very difficult. Based on geographic information system (GIS) remote sensing (RS) platforms, six impact factors, namely land surface temperature, fault density, Gutenberg–Liszt B value, formation combination entropy, distance to river aeromagnetic anomaly were selected. Through establishment certainty factor model (CF), weights entropy (ICF) evidence...

Category-level object pose estimation aims to predict the 6D and 3D metric size of objects from given categories. Due significant intra-class shape variations among different instances, existing methods have mainly focused on estimating dense correspondences between observed point clouds their canonical representations, i.e., normalized coordinate space (NOCS). Subsequently, a similarity transformation is applied recover size. Despite these efforts, current approaches still cannot fully...

The Hausdorff distance (HD) between two point sets is widely used in similarity measures, but the high computational cost of HD algorithms restrict their practical use. In this paper, we analyze time complexity to compute an accurate and find that reducing iterations inner loop significantly contributes average cost. Based on observation nearest neighbor (NN) breakpoint current suggests a higher probability break next loop, present novel efficient approach for computing based diffusion...

3D model retrieval is all along a difficult and hotspot in computer vision. Recently, the view-based methods make use of advanced convolutional neural networks which achieve excellent results. However, structural information was destroyed by projection relevance multiple perspectives not considered. In order to resolve this problem, manuscript analyzes process human observation models imitates recognition through combination recurrent networks. Our approach can convert between into...

Since the feature maps of deep neural networks were adopted to compute representation style and content information, transfer (NST) methods have sprung up like mushrooms. But existing ignore a fundamental fact that or an artistic image not only contains information but also information. And we find there may be conflict between content. Motivated by this idea, propose novel method, which adopts detail layer loss. To avoid potential conflicts loss loss, just abandon latter. The smooth base...

Abstract Electromagnetic Flowmeter is widely used in industrial production and life, because it characterized by high precision little environmental interference electromagnetic flow measurement. However, the commonly low frequency rectangular wave excitation greatly reduces measurement accuracy range of flowmeter due to in-phase interference, differential other factors. Double-frequency trapezoidal has advantages both excellent zero-point stability strong ability suppress noise slurry fluid...

A new optimization method of pile-anchor support for foundation pit based on BP neural network was been proposed and applied in engineering example. Uniform test can be used to construct study samples efficiently. is taken advantage build a prediction model predicting results large number random samples. Then, according the constraint condition criterions, best result screened out from results. Through an example, it showed that this efficient with good economic practical value.

there are many uncertainty factors in the design process of deep foundation pit engineering, such as soil parameters, loading, which make calculated displacement, settlement and safety factor have randomness uncertainty. This paper combines uniform (UD) with BP neural network. The UD structures random samples. Then, network trains samples corresponding lateral ground to get response relationship respectively. On this basis, probability density distribution each parameter is obtained by...