No abstract

Suppression of seismic random noise is one critical step in seismic data processing. In recent years, the outstanding ability of deep learning to denoise seismic data is impressive. Unsupervised Deep Image Prior (DIP) model has achieved promising denoising results without training label. However, during training, these models first learn the effective seismic events in the noisy data, and then pick up the random noise afterwards, i.e., overfitting. Thus, the practicability of DIP hinges on good early stopping (ES) that catches the potentially noise free seismic data. In this respect, the majority of DIP works only demonstrate potential of the models by showing the peak performance accessing the ground truth as reference, but provides no clue about how to operationally catch near-peak output without the ground truth. In this paper, we investigate the ES strategy in seismic data denoising using DIP method, which consistently detects the performance of reconstruction sequence by observing its running variance (VAR). The adopted ES method incurs low computational overhead. Numerical tests on 2D/3D synthetic and field data demonstrate that compared with other stopping criteria, the ES method exhibits superiority in suppressing random noise and preserves the effective signals better.

No abstract

In decision-making process, decision-makers may make different decisions because of their different experiences and knowledge. The abnormal preference value given by the biased decision-maker (the value that is too large or too small in the original data) may affect the decision result. To make the decision fair and objective, this paper combines the advantages of the power average (PA) operator and the Bonferroni mean (BM) operator to define the generalized fuzzy soft power Bonferroni mean (GFSPBM) operator and the generalized fuzzy soft weighted power Bonferroni mean (GFSWPBM) operator. The new operator not only considers the overall balance between data and information but also considers the possible interrelationships between attributes. The excellent properties and special cases of these ensemble operators are studied. On this basis, the idea of the bidirectional projection method based on the GFSWPBM operator is introduced, and a multi-attribute decision-making method, with a correlation between attributes, is proposed. The decision method proposed in this paper is applied to a software selection problem and compared to the existing methods to verify the effectiveness and feasibility of the proposed method.