Exploring the path of environmental visual design based on the perspective of new media

Abstract

Abstract This paper first constructs a convolutional neural network model by convolutional layer, activation layer and normalization layer and then uses the style migration algorithm and image semantic segmentation algorithm based on the convolutional neural network to construct an environmental visual design model based on the style migration algorithm. The experimental analysis of the effective application of the environmental visual design scheme under the new media perspective is conducted after data pre-processing. The results show that from the total loss function graph after weighting, the function decreases from 0.03 to 0.01 after a small fluctuation, and the style migration imaging tends to a stable state, and the model is optimal now. This demonstrates the exploration of environmental visual design and encourages the innovative development of visual design in the new media environment.