首页>
外国专利>
IMAGE RETRIEVAL METHOD BASED ON FEATURE EXTRACTION
IMAGE RETRIEVAL METHOD BASED ON FEATURE EXTRACTION
展开▼
机译：基于特征提取的图像检索方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
Disclosed is an image retrieval method based on feature extraction, applicable to determining at least one sample image similar to an image to be retrieved, the method comprising the following steps: a. performing a feature extraction step on all sample images to obtain all corresponding feature information, wherein the sample images are I1, I2, …, Is, and the feature information items corresponding to the sample images are V1, V2, …, Vs (S101); b. creating an index tree on the basis of all of the feature information (S102); c. acquiring the feature information corresponding to the image to be retrieved on the basis of the feature extraction step (S103); and d. retrieving the feature information corresponding to at least one sample image having the shortest distance on the basis of the index tree and the feature information corresponding to the image to be retrieved (S104). The method is powerful and simple, and can be used to prevent excessive influence of the image background on the image feature during processing and filtering of feature extraction, thus leading to excellent retrieval results and high commercial value.
展开▼
机译：公开了一种基于特征提取的图像检索方法，适用于确定与待检索图像相似的至少一个样本图像，该方法包括以下步骤：对所有样本图像进行特征提取步骤，得到所有对应的特征信息，其中，样本图像为I 1 ，I 2 ，…，I s ，与样本图像相对应的特征信息项为V 1 ，V 2 ，…，V s （S101）; b。基于所有特征信息创建索引树（S102）; C。在特征提取步骤的基础上，获取与待提取图像相对应的特征信息（S103）; d。基于索引树和与要检索的图像相对应的特征信息，检索与至少一个具有最短距离的样本图像相对应的特征信息（S104）。该方法功能强大，简单，可用于在特征提取的处理和过滤过程中，防止图像背景对图像特征的过度影响，从而获得了优异的检索结果，具有较高的商业价值。
展开▼
