Focusing on the road-railway level-crossings of the urban, basing on the attribute mathematics methods, the access safety influence factors of the railway level-crossings are analyzed. Basing on the analytic hierarchy process method, the assessment indexes, which impair the level-crossings system safety status, are set. Firstly, the attribute measurements of every single-index are calculated by mean of the construction of attribute measurement functions. Secondly, according to the similarity function theory, the similarity weights of the influence factors are defined, so as to assess the multi-indexes measurements objectively. Finally, make use of the confidence criterion to recognize the level-crossing system safety. Taking one of the railway level-crossings in Dalian as an example, the assessment results show that the attribute recognition result would not be affected by the index weight, namely, the weight values of the assessing indexes are considered as the same. The study shows that the method is feasible, reliable and easy for the synthetic safety assessment of the urban railway level-crossings safety
Improving Facial Attribute Recognition by Group and Graph Learning
