Pancreas segmentation is challenging due to the small proportion and highly changeable anatomical structure. It motivates us to propose a novel segmentation framework, namely Curriculum Knowledge Switching (CKS) framework, which decomposes detecting pancreas into three phases with different difficulty extent: straightforward, difficult, and challenging. The framework switches from straightforward to challenging phases and thereby gradually learns to detect pancreas. In addition, we adopt the momentum update parameter updating mechanism during switching, ensuring the loss converges gradually when the input dataset changes. Experimental results show that different neural network backbones with the CKS framework achieved state-of-the-art performance on the NIH dataset as measured by the DSC metric. The code is available at https://github.com/kunzhan/CKS_Pancreas

In practical engineering optimization problems (such as risk assessments), the parameters of the objective functions can be intervals because of noise and uncertainty; however, such problems cannot be solved by traditional multiobjective optimization methods. Yet, very little study has addressed interval multiobjective optimization methods compared to traditional multiobjective optimization methods. Therefore, a novel interval multiobjective optimization method called the Interval Cooperative Multiobjective Artificial Bee Colony Algorithm (ICMOABC) based on multiple populations for multiple objectives and interval credibility is proposed. Interval credibility is selected as the interval dominant method. Interval credibility is easy to combine with multiobjective optimization methods because it can describe the mean and width of intervals without increasing the dimension of the objective functions. The proposed algorithm has M single-objective optimization subpopulations updated by artificial bee colony algorithm, meaning it uses evolutionary resources more efficiently. In order to bring in diversity, the elitist learning strategy is used in the archive. The results of ICMOABC on various benchmark problems sets with different characteristics demonstrate its superior performance compared to some state-of-the-art algorithms.