Development And Application Of Human-machine Collaborative Product Modeling Bionic Design System Based On Deep Learnin

In recent years,with the rise of generative artificial intelligence such as AI painting and Chat GPT,the concept of product design has gradually begun to evolve from the traditional’manual production’ to human-centred ’intelligent generative design’.This new humanmachine collaborative approach is currently a major hotspot in product design,while traditional bionic design modelling methods are mostly designed by designers based on intuition and their own experience,with a certain degree of ambiguity and objectivity,easily causing the bionic product to be cognitively ambiguous,difficult to evoke emotional resonance,and difficult to meet the current designers’ design expression needs.In this study,the main focus is on the collision of artificial intelligence technology and product designers in the bionic design of inspiration and the process of solution generation,bionic design is originally an abstract process,if the abstract biological inspiration and the fusion process of figurative product shape can be visualised,and combined with artificial intelligence technology to express the designer’s creativity,it will greatly improve the efficiency of product shape bionic design and the accuracy rate,for This is of great significance to the innovative design and development of bionic products.This paper focuses on the following research on the intelligent bionic design of products with human-machine collaboration:(1)A deep learning-based bionic design approach for product modellingA deep learning-based bionic design method for product modelling is proposed to address problems such as the lack of a scientifically valid mapping strategy between bionic objects and bionic products in bionic fusion.The method combines bio-inspired design with a deep generative model,uses Style GAN’s image deformation technique to visualise the potential relationship between products and creatures,and generates new bionic fusion solutions.Experiments demonstrate that the fusion effect of the solutions generated by the method is significantly better than the fusion effect of feature point-based image deformation,and finally combines perceptual engineering and eye-movement experiments to jointly build a human-computer collaborative Finally,combined with perceptual engineering and eyemovement experiments,we jointly build a human-machine collaborative biologically inspired design model,which can accelerate the innovation and development of bionic products,and provide designers with design reference and rapid generation of bionic fusion solutions.(2)Deep learning-based sketch generation of real product bionic optimization methodsA deep learning-based sketch generation method for real products is proposed to address the lack of a collaborative human-computer approach that combines deep learning techniques with designers’ design expressions more effectively in bionic optimization.In order to allow designers to participate in the bionic design in a better human-machine collaborative way,a Style GAN-based sketch generation method for product design effect images is proposed,in which the sketches of design expressions can not only migrate the style of the target image more comprehensively,but also introduce certain deformation in the shape of the style migration to satisfy users with different degrees of design expressions.At the same time,to improve the quality of the generated new product images,similarity matching between sketches and style images is performed using perceptual engineering before style migration.Designers can use the sketches to express their bionic ideas,or choose the style of product samples that match the sketches by subjective matching or perceptual matching,and experimentally verify the effectiveness of this human-machine collaboration.(3)Research on product shape bionic evaluation method based on deep learningA deep learning-based product modeling bionic evaluation method is proposed to address the problems such as the lack of a scientific design evaluation method in bionic evaluation.The method is based on deep learning technology to design a shallow convolutional neural network model(S-VGG),in making the model training with high recognition accuracy,based on this condition,the probability that the image of a bionic product solution can be assigned to a biological class as the similarity between this image and the characteristics of the biological class,and use this as the basis for bionic evaluation,helping designers to extract from a large number of bionic solutions the most bionic value of the image,forming a more scientific,objective and reliable evaluation system.(4)Human-machine collaboration for product bionic design applicationsThe above theoretical models are integrated to develop a collaborative human-machine product bionic design aid prototyping system,which provides designers with inspiration for bionic design and a rapid method for generating solutions,as well as a scientific and reliable method for evaluating bionic solutions.The system mainly takes bionic matching,bionic fusion,bionic optimisation and bionic matching as the core of system construction,and validates its application through actual product bionic design.