Design of Flower Recognition System Based on PyTorch

With the rapid growth of the Chinese economy and the improvement of living standards, flowers have become widely popular decorations and gifts. This design developed a flower recognition system founded on PyTorch to attain efficient and precise flower recognition. The Oxford-102 flower dataset offered by the University of Oxford was utilized for model training, which encompasses 102 categories of flower images, totaling 7,370 images. 88% of the images were chosen as the training set, while the remaining 12% were designated as the validation set. Image preprocessing and image enhancement techniques were employed to enhance the generalization capability of the model. A ResNet residual network was adopted to construct the classification model, and it was trained and verified using transfer learning approaches. The experimental outcomes demonstrated that the model achieved an accuracy rate of 93.41% in the flower recognition task, and API development was carried out to build a flower recognition system. This study indicates that machine vision technology can accomplish efficient and precise flower recognition, which will bring convenience to flower enthusiasts, gardeners, and even promote the development of the flower industry.