Chest X-ray images are widely used in clinical practice such as diagnosis and treatment. The automatic radiology report generation system can effectively reduce the rate of misdiagnosis and missed diagnosis. Previous studies were focused on the long text generation problem of image paragraph, ignoring the characteristics of the image and the auxiliary role of patient background information for diagnosis. In this paper, we propose a new hierarchical model with multi-attention considering the background information. The multi-attention mechanism can focus on the image's channel and spatial information simultaneously, and map it to the sentence topic. The patient's background information will be encoded by the neural network first, then it will be aggregated into a vector representation by a multi-layer perception and added to the pretrained vanilla word embedding, which finally forms a new word embedding after fusion. Our experimental results demonstrated that the model outperforms all baselines, achieving the state-of-the-art performance in terms of accuracy.
Isothermal, strain-controlled low cycle fatigue (LCF) tests of P92 steel are conducted in air at different temperatures ranging from 20 to 650 8C and various strain ranges to clarify the influence of temperature on its LCF behavior. A symmetrical triangular wave form and a constant 1.0 Â 10 À3 s À1 strain rate are employed for all the tests. Cyclic softening behavior is expressed in two ways: peak tensile stress in terms of cycle numbers and in terms of accumulated inelastic strain. Based on the definition of "inelastic ratio," correlation between the value of accumulated inelastic strain ( e in;II ) at the beginning of the second softening stage and the inelastic ratio can be distinctively divided into three regions. The modified cyclic plasticity model, in which a linear term is added into the isotropic hardening variable, is developed to simulate the stress-strain hysteresis loop and the cyclic softening curve. Additionally, parameters in the modified isotropic hardening variable at different temperatures are correlated to the cyclic softening phenomenon. Performances of the modified model are assessed by comparing predictions with experimental results. Good predictive capability of the modified constitutive model is obtained.[ Ã ] X.
While a certain correlation between the cation exchange capacity (CEC) of the soil and the strength of the cement stabilized soil has been reported, the mechanism remains unclear. In this research, a set of soil samples with different CECs were stabilized with different proportions of cement and calcium hydroxide (Ca(OH) 2 , CH). The influence of soil CEC on the strength of the stabilized soil was investigated by analyzing the CH saturation in the pore solution and measuring the strength of the stabilized soil specimens. It is revealed that cation exchange in the soil can reduce the CH saturation of the stabilized soil. If the CEC of the soil is too high, the CH in the pore solution of the stabilized soil cannot reach the saturation level, and further cation exchange would then consume the Ca 2 þ ions which should be originally used to generate calcium silicate hydrate, thus result in the poor strength of the stabilized soil.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.