“酱”系列酒包装设计风格-酱艺图像集合

本数据采用“酱酒工艺+非遗艺术+视觉风格”的三级标注规则，酱酒工艺标签区分制曲、发酵、蒸馏等酿造工序，非遗艺术标签提取陶艺、竹编、篆刻等非遗形式，视觉风格标签划分新中式艺术、传统工艺写实、抽象工艺表达等类型。算法层面，基于ResNet-101模型搭建图像特征识别框架，通过深度卷积运算捕捉包装图像的工艺艺术细节与视觉特征，实现标签自动匹配，经测试，特征识别准确率达95%，可高效完成数据分类与检索。This dataset adopts a three-level annotation framework of "Jiangjiu Brewing Craft + Intangible Cultural Heritage Art + Visual Style". The Jiangjiu brewing craft labels distinguish core brewing procedures including starter making, fermentation, and distillation. The intangible cultural heritage art labels cover representative intangible cultural heritage forms such as pottery art, bamboo weaving, and seal engraving. The visual style labels classify packaging images into categories like neo-Chinese art, traditional craft realism, and abstract craft expression. From an algorithmic standpoint, an image feature recognition framework is constructed based on the ResNet-101 model. Deep convolutional operations are employed to capture craft art details and visual features of packaging images, enabling automatic label matching. Experimental tests have verified that the feature recognition accuracy reaches 95%, allowing efficient data classification and retrieval.

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于酱酒包装设计风格的图像集合，由贵州酱酒集团有限公司自行产生，规模为10G，每周更新。它采用三级标注规则结合ResNet-101算法，准确率达95%，主要用于支持酱酒包装的艺术设计、教学研究及文创开发，强调将酿造工艺与非遗艺术融入视觉表达。