However, conventional BP neural networks have many short-comings, such as slow convergence, sensitivity to initial values andeasily getting into local minimum. Meanwhile, the color correctionmodel generated by the BP neural network method is not unique. Asa result, this paper brings about improvements to traditional neu-ral network by adopting the mind evolutionary computation (MEC)and adaboost algorithm (Adaboost) to optimize the BP neural net-work, and hence proposes the MEC-BP-Adaboost neural network.The proposed algorithm can effectively avoid BP neural networkgetting into local minimum and achieve relevant results after eachtraining