CN111914921A - A method and system for similarity image retrieval based on multi-feature fusion - Google Patents
A method and system for similarity image retrieval based on multi-feature fusionDownload PDFInfo
- Publication number
- CN111914921A CN111914921ACN202010725870.0ACN202010725870ACN111914921ACN 111914921 ACN111914921 ACN 111914921ACN 202010725870 ACN202010725870 ACN 202010725870ACN 111914921 ACN111914921 ACN 111914921A
- Authority
- CN
- China
- Prior art keywords
- image
- similarity
- feature
- feature fusion
- image retrieval
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本发明涉及图像识别领域，具体而言，涉及一种基于多特征融合的相似性图像检索方法及系统。The invention relates to the field of image recognition, in particular, to a method and system for similarity image retrieval based on multi-feature fusion.
背景技术Background technique
随着数字媒体技术的广泛应用，海量的图像已经成为了生活中必不可少的部分，在教育、文化、生命科学等多个领域有非常广泛的应用。给定一张特定的图像，如果可以从海量的图像找到它的相似图像有非常好的实际应用价值。With the wide application of digital media technology, massive images have become an indispensable part of life, and have a very wide range of applications in education, culture, life sciences and other fields. Given a specific image, it has very good practical application value if it can find its similar images from a large number of images.
传统的图像检索方法往往高度依赖于训练样本，导致了经典的方法往往不能直接应用于单幅相似图像的检索，具体存在以下的缺陷：Traditional image retrieval methods are often highly dependent on training samples, which leads to the fact that classical methods cannot be directly applied to the retrieval of a single similar image, with the following defects:
1.无法充分地提取和融合一幅图像的全局特征和局部特征，并有效地进行融合；1. The global features and local features of an image cannot be fully extracted and fused, and the fusion is performed effectively;
2.不能够有效地度量不同图像之间的相似度。2. The similarity between different images cannot be effectively measured.
发明内容SUMMARY OF THE INVENTION
本发明的目的在于提供一种基于多特征融合的相似性图像检索方法，其能够利用SIFT和GIST特征算子相结合的方式，更加充分地提取了图像的局部和全局特征，并进行了较为充分的融合。The purpose of the present invention is to provide a similarity image retrieval method based on multi-feature fusion, which can use the combination of SIFT and GIST feature operators to more fully extract the local and global features of the image, and perform more adequate fusion.
本发明的另一目的在于提供一种基于多特征融合的相似性图像检索系统，其能够运行一种基于多特征融合的相似性图像检索方法。Another object of the present invention is to provide a similarity image retrieval system based on multi-feature fusion, which can run a similarity image retrieval method based on multi-feature fusion.
本发明的实施例是这样实现的：Embodiments of the present invention are implemented as follows:
第一方面，本申请实施例提供一种基于多特征融合的相似性图像检索方法，其包括如下步骤：通过GIST特征算子对图像进行全局特征提取；通过SIFT特征算子对图像进行局部特征提取；计算不同图像之间的相似度；判断相似度是否大于第一预设阈值，若是，则判定检索出的图像为相似图像，若不是，则删除检索出的图像。In a first aspect, an embodiment of the present application provides a method for retrieving similarity images based on multi-feature fusion, which includes the following steps: extracting global features of an image by using a GIST feature operator; extracting local features of an image by using a SIFT feature operator Calculate the similarity between different images; determine whether the similarity is greater than the first preset threshold, if so, determine that the retrieved image is a similar image, if not, delete the retrieved image.
在本发明的一些实施例中，上述通过SIFT特征算子对图像进行局部特征提取之后还包括通过BoW模型获得表征图像直方图。In some embodiments of the present invention, after performing local feature extraction on the image by using the SIFT feature operator, the method further includes obtaining a histogram representing the image by using the BoW model.
在本发明的一些实施例中，上述通过SIFT特征算子对图像进行局部特征提取包括通过SIFT特征算子从图像中提取视觉词汇向量。In some embodiments of the present invention, performing local feature extraction on the image by using the SIFT feature operator includes extracting a visual vocabulary vector from the image by using the SIFT feature operator.
在本发明的一些实施例中，上述还包括通过K-means对词义相近的视觉词汇进行合并，构造一个包含K个词汇的单词表。In some embodiments of the present invention, the above also includes combining visual words with similar meanings through K-means to construct a word list containing K words.
在本发明的一些实施例中，上述还包括对每个单词在图像中出现的次数进行统计，并将图像表征为一个K维数值向量。In some embodiments of the present invention, the above further includes counting the number of times each word appears in the image, and representing the image as a K-dimensional numerical vector.
在本发明的一些实施例中，上述计算不同图像之间的相似度包括根据GIST特征算子提取特征的基础，通过欧式距离对不同图像的全局相似度进行计算。In some embodiments of the present invention, the above calculation of the similarity between different images includes calculating the global similarity of different images through Euclidean distance on the basis of extracting features according to the GIST feature operator.
在本发明的一些实施例中，上述还包括根据表征直方图的基础，通过巴氏距离对不同图像的局部相似度进行计算。In some embodiments of the present invention, the above-mentioned method further includes calculating the local similarity of different images through the Bavarian distance based on the representation histogram.
在本发明的一些实施例中，上述还包括将图像的局部相似度和全局相似度结合，作为图像相似性检索的依据。In some embodiments of the present invention, the above-mentioned method further includes combining the local similarity and the global similarity of the images as the basis for the similarity retrieval of the images.
第二方面，本申请实施例提供一种基于多特征融合的相似性图像检索系统，包括第一提取模块，用于通过GIST特征算子对图像进行全局特征提取；第二提取模块，用于通过SIFT特征算子对图像进行局部特征提取；计算模块，用于计算不同图像之间的相似度；判断模块，用于判断相似度是否大于第一预设阈值，若是，则判定检索出的图像为相似图像，若不是，则删除检索出的图像。In a second aspect, an embodiment of the present application provides a similarity image retrieval system based on multi-feature fusion, including a first extraction module for extracting global features of an image by using a GIST feature operator; a second extraction module for performing global feature extraction on images by The SIFT feature operator extracts local features of the image; the calculation module is used to calculate the similarity between different images; the judgment module is used to judge whether the similarity is greater than the first preset threshold, and if so, the retrieved image is judged to be Similar images, if not, delete the retrieved images.
在本发明的一些实施例中，上述还包括用于存储计算机指令的至少一个存储器；与存储器通讯的至少一个处理器，其中当至少一个处理器执行计算机指令时，至少一个处理器使系统执行：第一提取模块、第二提取模块、计算模块以及判断模块。In some embodiments of the invention, the above also includes at least one memory for storing computer instructions; at least one processor in communication with the memory, wherein when the at least one processor executes the computer instructions, the at least one processor causes the system to perform: A first extraction module, a second extraction module, a calculation module and a judgment module.
相对于现有技术，本发明的实施例至少具有如下优点或有益效果：Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:
其能够利用SIFT和GIST特征算子相结合的方式，更加充分地提取了图像的局部和全局特征，并进行了较为充分的融合，即使改变旋转角度，图像亮度或拍摄视角，仍然能够得到好的检测效果；利用自适应权重的方式，辅助任务将和主任务共享部分特征表示，在训练辅助任务的同时可以训练良好的特征表示，该特征表示将有利于加快主任务的学习，进而在相似度计算的过程中，可以考虑不同方法。It can use the combination of SIFT and GIST feature operators to more fully extract the local and global features of the image, and carry out a relatively sufficient fusion, even if the rotation angle, image brightness or shooting angle of view are changed, it can still get good results. Detection effect; using the adaptive weight method, the auxiliary task will share part of the feature representation with the main task, and a good feature representation can be trained while training the auxiliary task. In the calculation process, different methods can be considered.
附图说明Description of drawings
为了更清楚地说明本发明实施例的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本发明的某些实施例，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他相关的附图。In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore do not It should be regarded as a limitation of the scope, and for those of ordinary skill in the art, other related drawings can also be obtained according to these drawings without any creative effort.
图1为本发明实施例提供的一种基于多特征融合的相似性图像检索方法步骤示意图；1 is a schematic diagram of steps of a method for retrieving similarity images based on multi-feature fusion provided by an embodiment of the present invention;
图2为本发明实施例提供的一种基于多特征融合的相似性图像检索方法详细步骤示意图；2 is a schematic diagram of detailed steps of a method for retrieving similarity images based on multi-feature fusion provided by an embodiment of the present invention;
图3为本发明实施例提供的一种基于多特征融合的相似性图像检索系统模块示意图。FIG. 3 is a schematic diagram of modules of a similarity image retrieval system based on multi-feature fusion according to an embodiment of the present invention.
具体实施方式Detailed ways
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。通常在此处附图中描述和示出的本申请实施例的组件可以以各种不同的配置来布置和设计。In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments It is a part of the embodiments of the present application, but not all of the embodiments. The components of the embodiments of the present application generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations.
因此，以下对在附图中提供的本申请的实施例的详细描述并非旨在限制要求保护的本申请的范围，而是仅仅表示本申请的选定实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。Thus, the following detailed description of the embodiments of the application provided in the accompanying drawings is not intended to limit the scope of the application as claimed, but is merely representative of selected embodiments of the application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.
下面结合附图，对本申请的一些实施方式作详细说明。在不冲突的情况下，下述的各个实施例及实施例中的各个特征可以相互组合。Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The various embodiments described below and various features of the embodiments may be combined with each other without conflict.
实施例1Example 1
请参阅图1，图1为本发明实施例提供的一种基于多特征融合的相似性图像检索方法步骤示意图，其包括如下步骤：Please refer to FIG. 1. FIG. 1 is a schematic diagram of steps of a similarity image retrieval method based on multi-feature fusion provided by an embodiment of the present invention, which includes the following steps:
步骤S100，通过GIST特征算子对图像进行全局特征提取；Step S100, perform global feature extraction on the image through the GIST feature operator;
具体的，通过GIST算子降采样进行场景级别的识别对图像进行全局特征提取。Specifically, the scene-level recognition is performed through the down-sampling of the GIST operator to extract the global features of the image.
在一些实施方式中，GIST描述本身就是一幅图像的整体信息，适用于场景分类与分析，例如，GIST算子对高分辨率遥感影像进行识别分类，类似于降采样，通俗一点地讲就是场景级别的识别，比如选取任意一张谷歌地图的局部部分，可以对图片划分成多个正方形，每个正方形为300×300pix，然后用不同的颜色进行区分，可以是橙黄色是密集住宅区，绿色为水体等。In some embodiments, the GIST description itself is the overall information of an image, which is suitable for scene classification and analysis. For example, the GIST operator can identify and classify high-resolution remote sensing images, which is similar to downsampling. Level recognition, such as selecting a part of any Google map, you can divide the picture into multiple squares, each square is 300 × 300pix, and then use different colors to distinguish, it can be orange for dense residential areas, green for water bodies, etc.
步骤S110，通过SIFT特征算子对图像进行局部特征提取；Step S110, performing local feature extraction on the image through the SIFT feature operator;
具体的，通过SIFT求一幅图中的特征点及其有关scale和orientation的描述子得到特征并进行图像特征点匹配对图像进行局部特征提取，描述一块区域，使其具有高可区分度。Specifically, through SIFT, the feature points in a picture and the descriptors related to scale and orientation are obtained to obtain features, and image feature point matching is performed to extract local features of the image, describe an area, and make it highly distinguishable.
在一些实施方式中，通过SIFT在不同的尺度空间上查找关键点，也就是特征点，并计算出关键点的方向。采用SIFT可解决：目标的旋转、缩放、平移(RST)、图像仿射/投影变换(视点viewpoint)、光照影响(illumination)、目标遮挡(occlusion)、杂物场景(clutter)以及噪声等问题。解决目标的自身状态、场景所处的环境和成像器材的成像特性等因素影响图像配准/目标识别跟踪的性能。In some embodiments, key points, that is, feature points, are searched in different scale spaces through SIFT, and the directions of the key points are calculated. Using SIFT can solve: target rotation, scaling, translation (RST), image affine/projection transformation (viewpoint), illumination, target occlusion (occlusion), clutter and noise. It is solved that factors such as the state of the target, the environment in which the scene is located, and the imaging characteristics of the imaging equipment affect the performance of image registration/target recognition and tracking.
步骤S120，计算不同图像之间的相似度；Step S120, calculating the similarity between different images;
在一些实施方式中，可以采用余弦相似度计算，把图片表示成一个向量，通过计算向量之间的余弦距离来表征两张图片的相似度；也可以采用哈希算法计算图片的相似度，包括aHash、pHash、dHash。感知哈希不是以严格的方式计算Hash值，而是以更加相对的方式计算哈希值，因为“相似”与否，就是一种相对的判定。值哈希算法、差值哈希算法和感知哈希算法都是值越小，相似度越高，取值为0-64，即汉明距离中，64位的hash值有多少不同。三直方图和单通道直方图的值为0-1，值越大，相似度越高；还可以采用直方图计算图片的相似度，按照颜色的全局分布情况来看待，对于灰度图可以将图片进行等分，然后在计算图片的相似度。In some embodiments, the cosine similarity calculation can be used to represent the picture as a vector, and the similarity between the two pictures can be represented by calculating the cosine distance between the vectors; a hash algorithm can also be used to calculate the similarity of the pictures, including aHash, pHash, dHash. Perceptual hashing does not calculate the hash value in a strict way, but calculates the hash value in a more relative way, because "similar" or not is a relative judgment. The value hash algorithm, the difference value hash algorithm and the perceptual hash algorithm are all the smaller the value, the higher the similarity, and the value is 0-64, that is, how different the 64-bit hash value is in the Hamming distance. The value of three histograms and single-channel histograms is 0-1, the larger the value, the higher the similarity; the histogram can also be used to calculate the similarity of the picture, according to the global distribution of colors, for grayscale images, you can use The pictures are divided into equal parts, and then the similarity of the pictures is calculated.
步骤S130，判断相似度是否大于第一预设阈值；Step S130, judging whether the similarity is greater than a first preset threshold;
具体的，判断相似度是否大于第一预设阈值，若相似度大于第一预设阈值，则进入步骤S140，若相似度小于等于第一预设阈值，则进入步骤S150。Specifically, it is judged whether the similarity is greater than the first preset threshold, if the similarity is greater than the first preset threshold, go to step S140, and if the similarity is less than or equal to the first preset threshold, go to step S150.
在一些实施方式中，第一预设阈值可以是根据图片不同而进行变化的，例如，假定一张图片共有n个像素，其中灰度值小于阈值的像素为n1个，大于等于阈值的像素为n2个(n1+n2＝n)。w1和w2表示这两种像素各自的比重，而所有灰度值小于阈值的像素的平均值和方差分别为μ1和σ1，所有灰度值大于等于阈值的像素的平均值和方差分别为μ2和σ2，对图像求直方图，用一系列从小到大的阀值去取一下，分别代入BBS的算式。使得“类内差异最小”或“类间差异最大”的那个值，就是最终的阈值，例如，可以是0.4、0.5、0.6等。In some embodiments, the first preset threshold may be changed according to different pictures. For example, it is assumed that a picture has a total of n pixels, wherein the pixels whose gray value is less than the threshold are n1, and the pixels whose gray value is greater than or equal to the threshold are n2 (n1+n2=n). w1 and w2 represent the respective proportions of these two types of pixels, and the average and variance of all pixels with gray values less than the threshold are μ1 and σ1, respectively, and the average and variance of all pixels with gray values greater than or equal to the threshold are μ2 and σ2, find the histogram of the image, use a series of thresholds from small to large to take them, and substitute them into the BBS formula. The value that makes "the smallest difference between classes" or "the largest difference between classes" is the final threshold, for example, it can be 0.4, 0.5, 0.6 and so on.
步骤S140，检索出的图像为相似图像；Step S140, the retrieved images are similar images;
具体的，将检索出的图片判定为相似图像。Specifically, the retrieved pictures are determined as similar images.
步骤S150，删除检索出的图像。Step S150, delete the retrieved image.
实施例2Example 2
请参阅图2，图2为本发明实施例提供的一种基于多特征融合的相似性图像检索方法详细步骤示意图，其包括如下步骤：Please refer to FIG. 2. FIG. 2 is a schematic diagram of the detailed steps of a similarity image retrieval method based on multi-feature fusion provided by an embodiment of the present invention, which includes the following steps:
步骤S200，通过GIST特征算子对图像进行全局特征提取；Step S200, perform global feature extraction on the image through the GIST feature operator;
具体的，通过GIST算子降采样进行场景级别的识别对图像进行全局特征提取。可参考图1步骤S100的相关描述，这里不再赘述。Specifically, the scene-level recognition is performed through the down-sampling of the GIST operator to extract the global features of the image. Reference may be made to the relevant description of step S100 in FIG. 1 , which will not be repeated here.
步骤S210,通过SIFT特征算子从图像中提取视觉词汇向量；Step S210, extracts visual vocabulary vector from image by SIFT feature operator;
具体的，通过SIFT进行特征提取用来从图片中提取出关键点、或特征点、角点、视觉词汇向量。Specifically, feature extraction through SIFT is used to extract key points, or feature points, corner points, and visual vocabulary vectors from the image.
在一些实施方式中，特征提取可以包括：关键点的位置(坐标)；关键点的尺度，如果没有提取出尺度信息，则该算法在对特征点描述时候将不能依据尺度信息进行描述，进而该算法不具有尺度不变性；关键点的方向，如果没有提取出方向信息，则该算法在对特征点描述时候将不能依据方向信息进行描述，进而该算法不具有旋转不变性。有了关键点的信息，之后将对关键点进行描述，从而可以根据不同关键点的不同描述来判断关键点之间的匹配关系。In some embodiments, the feature extraction may include: the position (coordinate) of the key point; the scale of the key point, if the scale information is not extracted, the algorithm will not be able to describe the feature point according to the scale information, and then the The algorithm does not have scale invariance; if the direction of key points is not extracted, the algorithm will not be able to describe the feature points according to the direction information, and the algorithm has no rotation invariance. With the information of the key points, the key points will be described later, so that the matching relationship between the key points can be judged according to different descriptions of different key points.
步骤S220,通过K-means对词义相近的视觉词汇进行合并，构造一个包含K个词汇的单词表；Step S220, by K-means, the visual vocabulary with similar meanings is merged, and a word list containing K vocabulary is constructed;
具体的，K-means是基于欧式距离的聚类算法，其认为两个目标的距离越近，相似度越大，对词义相近的视觉词汇进行合并，构造一个包含K个词汇的单词表。Specifically, K-means is a clustering algorithm based on Euclidean distance. It considers that the closer the distance between the two targets, the greater the similarity. The visual words with similar meanings are merged to construct a word list containing K words.
在一些实施方式中，基于欧式距离的K-means假设了各个数据簇的数据具有一样的的先验概率并呈现球形分布，但这种分布在实际生活中并不常见。面对非凸的数据分布形状时我们可以引入核函数来优化，这时算法又称为核K-means算法，是核聚类方法的一种。核聚类方法的主要思想是通过一个非线性映射，将输入空间中的数据点映射到高位的特征空间中，并在新的特征空间中进行聚类。非线性映射增加了数据点线性可分的概率，从而在经典的聚类算法失效的情况下，通过引入核函数可以达到更为准确的聚类结果。In some embodiments, K-means based on Euclidean distance assumes that the data of each data cluster has the same prior probability and exhibits a spherical distribution, but this distribution is not common in real life. In the face of non-convex data distribution shape, we can introduce kernel function to optimize. At this time, the algorithm is also called kernel K-means algorithm, which is a kind of kernel clustering method. The main idea of the kernel clustering method is to map the data points in the input space to the high-level feature space through a nonlinear mapping, and perform clustering in the new feature space. The nonlinear mapping increases the probability that the data points are linearly separable, so that when the classical clustering algorithm fails, more accurate clustering results can be achieved by introducing a kernel function.
步骤S230,对每个单词在图像中出现的次数进行统计，并将图像表征为一个K维数值向量；Step S230, counts the number of times that each word appears in the image, and characterizes the image as a K-dimensional numerical vector;
在一些实施方式中，聚类的效果不单单是从数据的数量和维度，更多的要在dataexploration阶段对所需聚类的数据进行合理的处理和筛选，以及数据本身的意义出发。K-means是用的最广的一种方法，不需要相关的模型assumption同时有较好的聚类结果，如此万金油的方法在处理特定问题时可能并不是最优选择，通常在聚类分析时会同时运用几种不同的方法来比较其效果和解释程度，还可以采用层次聚类和混合模型。In some embodiments, the effect of clustering depends not only on the quantity and dimensions of the data, but also on the reasonable processing and screening of the data to be clustered in the dataexploration stage, as well as the meaning of the data itself. K-means is the most widely used method. It does not require related model assumptions and has good clustering results. Such a panacea method may not be the best choice when dealing with specific problems, usually in cluster analysis. Several different methods are used simultaneously to compare their effect and degree of interpretation, and hierarchical clustering and mixed models can also be used.
步骤S240,通过BoW模型获得表征图像直方图；Step S240, obtaining a characteristic image histogram through the BoW model;
具体的，通过Bag Of Word模型，用于图像检索的一种方法，获得表征图像直方图。Specifically, a histogram representing an image is obtained through the Bag Of Word model, a method used for image retrieval.
在一些实施方式中，Bag Of Features与Bag Of Word原理类似，但适用于图像的检索，具体而言，其可以分为以下步骤：In some embodiments, Bag Of Features is similar in principle to Bag Of Word, but is suitable for image retrieval. Specifically, it can be divided into the following steps:
特征提取：将图像看成一个由各种图像块组成的集合，通过特征提取，获得图像的关键图像特征；Feature extraction: treat the image as a collection of various image blocks, and obtain the key image features of the image through feature extraction;
学习“视觉词典”(visual vocabulary)：获得了多张图像的特征之后，这些特征并没有通过分类处理，其中有的特征点之间是极其相似，所以这一步骤通过K-means聚类算法，将我们提取出来的特征点进行分类处理。聚类是学习视觉词典的重点操作。将聚类出来的聚类中心称为视觉单词。而将视觉单词组成的集合称为视觉词典/码本；Learning "visual vocabulary": After obtaining the features of multiple images, these features have not been classified and processed, and some of the feature points are extremely similar, so this step passes the K-means clustering algorithm, The feature points we extracted are classified. Clustering is the key operation for learning visual dictionaries. The clustered centers are called visual words. And the collection of visual words is called visual dictionary/codebook;
对输入特征集进行量化：将输入的特征集合，映射到上一步做来的码本之中。通过计算输入特征到视觉单词的距离，然后将其映射到距离最近的视觉单词中，并计数；Quantize the input feature set: map the input feature set to the codebook from the previous step. By calculating the distance from the input feature to the visual word, then mapping it to the nearest visual word, and counting;
把输入图像转化成视觉单词(visual words)的频率直方图：这一步骤通过对图像特征提取，然后将提取出来的特征点，根据上一步，转换为频率直方图。Convert the input image into a frequency histogram of visual words: This step extracts image features, and then converts the extracted feature points into a frequency histogram according to the previous step.
步骤S250,根据GIST特征算子提取特征的基础，通过欧式距离对不同图像的全局相似度进行计算；Step S250, calculates the global similarity of different images by Euclidean distance on the basis of extracting features according to the GIST feature operator;
在一些实施方式中，通过欧氏距离，衡量个体在空间上存在的距离，距离越远说明个体间的差异越大，衡量的是n维空间中两个点之间的实际距离。In some embodiments, the distance between individuals in space is measured by Euclidean distance. The farther the distance is, the greater the difference between individuals is, and the actual distance between two points in the n-dimensional space is measured.
步骤S260,根据表征直方图的基础，通过马氏距离对不同图像的局部相似度进行计算；Step S260, according to the basis of the representation histogram, calculate the local similarity of different images by Mahalanobis distance;
具体的，通过马氏距离对欧氏距离进行修正，修正欧氏距离中各个维度尺度不一致且相关的问题，对不同图像的局部相似度进行计算。Specifically, the Euclidean distance is corrected by the Mahalanobis distance, and the problem of inconsistent and related dimension scales in the Euclidean distance is corrected, and the local similarity of different images is calculated.
步骤S270,将图像的局部相似度和全局相似度结合，作为图像相似性检索的依据；Step S270, combining the local similarity and the global similarity of the image as a basis for image similarity retrieval;
具体的，对基准图像和数据库中的一个或多个图像都进行多特征深度提取，得到基准表征，在特征提取的基础上，利用相似度计算准则对不同图像间的相似度进行计算，然后将图像的局部相似度和全局相似度结合，作为图像相似性检索的依据。Specifically, multi-feature depth extraction is performed on the reference image and one or more images in the database to obtain the reference representation. On the basis of the feature extraction, the similarity between different images is calculated using the similarity calculation criterion, and then the The local similarity and global similarity of images are combined as the basis for image similarity retrieval.
步骤S280,判断相似度是否大于第一预设阈值；Step S280, judging whether the similarity is greater than the first preset threshold;
具体的，判断相似度是否大于第一预设阈值，若相似度大于第一预设阈值，则进入步骤S290，若相似度小于等于第一预设阈值，则进入步骤S300。Specifically, it is determined whether the similarity is greater than the first preset threshold, and if the similarity is greater than the first preset threshold, the process proceeds to step S290, and if the similarity is less than or equal to the first preset threshold, the process proceeds to step S300.
在一些实施方式中，第一预设阈值可以是根据图片不同而进行变化的，例如，假定一张图片共有n个像素，其中灰度值小于阈值的像素为n1个，大于等于阈值的像素为n2个(n1+n2＝n)。w1和w2表示这两种像素各自的比重，而所有灰度值小于阈值的像素的平均值和方差分别为μ1和σ1，所有灰度值大于等于阈值的像素的平均值和方差分别为μ2和σ2，对图像求直方图，用一系列从小到大的阀值去取一下，分别代入BBS的算式。使得“类内差异最小”或“类间差异最大”的那个值，就是最终的阈值，例如，可以是0.4、0.5、0.6等。In some embodiments, the first preset threshold may be changed according to different pictures. For example, it is assumed that a picture has a total of n pixels, wherein the pixels whose gray value is less than the threshold are n1, and the pixels whose gray value is greater than or equal to the threshold are n2 (n1+n2=n). w1 and w2 represent the respective proportions of these two types of pixels, and the average and variance of all pixels with gray values less than the threshold are μ1 and σ1, respectively, and the average and variance of all pixels with gray values greater than or equal to the threshold are μ2 and σ2, find the histogram of the image, use a series of thresholds from small to large to take them, and substitute them into the BBS formula. The value that makes "the smallest difference between classes" or "the largest difference between classes" is the final threshold, for example, it can be 0.4, 0.5, 0.6 and so on.
步骤S290,检索出的图像为相似图像；Step S290, the retrieved images are similar images;
具体的，将检索出的图片判定为相似图像。Specifically, the retrieved pictures are determined as similar images.
步骤S300,删除检索出的图像。Step S300, delete the retrieved image.
实施例3Example 3
请参阅图3，图3为本发明实施例提供的一种基于多特征融合的相似性图像检索系统模块示意图。Please refer to FIG. 3 , which is a schematic diagram of modules of a similarity image retrieval system based on multi-feature fusion provided by an embodiment of the present invention.
一种基于多特征融合的相似性图像检索系统包括：第一提取模块，用于通过GIST特征算子对图像进行全局特征提取；第二提取模块，用于通过SIFT特征算子对图像进行局部特征提取；计算模块，用于计算不同图像之间的相似度；判断模块，用于判断相似度是否大于第一预设阈值，若是，则判定检索出的图像为相似图像，若不是，则删除检索出的图像。A similarity image retrieval system based on multi-feature fusion includes: a first extraction module for extracting global features of an image by using a GIST feature operator; a second extraction module for performing local feature extraction on an image by using a SIFT feature operator Extraction; a calculation module, used to calculate the similarity between different images; a judgment module, used to judge whether the similarity is greater than the first preset threshold, if so, determine that the retrieved image is a similar image, if not, delete the retrieval out image.
还包括存储器、处理器和通信接口，该存储器、处理器和通信接口相互之间直接或间接地电性连接，以实现数据的传输或交互。例如，这些元件相互之间可通过一条或多条通讯总线或信号线实现电性连接。存储器可用于存储软件程序及模块，处理器通过执行存储在存储器内的软件程序及模块，从而执行各种功能应用以及数据处理。该通信接口可用于与其他节点设备进行信令或数据的通信。It also includes a memory, a processor and a communication interface, the memory, the processor and the communication interface are electrically connected to each other directly or indirectly to realize the transmission or interaction of data. For example, these elements may be electrically connected to each other through one or more communication buses or signal lines. The memory can be used to store software programs and modules, and the processor executes various functional applications and data processing by executing the software programs and modules stored in the memory. The communication interface can be used for signaling or data communication with other node devices.
其中，存储器可以是但不限于，随机存取存储器(Random Access Memory，RAM)，只读存储器(Read Only Memory，ROM)，可编程只读存储器(Programmable Read-OnlyMemory，PROM)，可擦除只读存储器(Erasable Programmable Read-Only Memory，EPROM)，电可擦除只读存储器(Electric Erasable Programmable Read-Only Memory，EEPROM)等。Wherein, the memory may be, but not limited to, random access memory (Random Access Memory, RAM), read only memory (Read Only Memory, ROM), programmable read only memory (Programmable Read-Only Memory, PROM), erasable only memory Read memory (Erasable Programmable Read-Only Memory, EPROM), Electric Erasable Programmable Read-Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.
处理器可以是一种集成电路芯片，具有信号处理能力。该处理器102可以是通用处理器，包括中央处理器(Central Processing Unit，CPU)、网络处理器(NetworkProcessor，NP)等；还可以是数字信号处理器(Digital Signal Processing，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field－Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。The processor may be an integrated circuit chip with signal processing capabilities. The processor 102 may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; it may also be a digital signal processor (Digital Signal Processing, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
可以理解，图1图2图3所示仅为示意，还可包括比图1图2图3中所示更多或者更少的组件，或者具有与图1图2图3所示不同的配置。图3中所示的各组件可以采用硬件、软件或其组合实现。It can be understood that, those shown in FIGS. 1 , 2 and 3 are only schematic, and may further include more or less components than those shown in FIGS. 1 , 2 and 3 , or have different configurations from those shown in FIGS. 1 , 2 and 3 . . Each component shown in FIG. 3 can be implemented in hardware, software, or a combination thereof.
在本申请所提供的实施例中，应该理解到，所揭露的装置和方法，也可以通过其它的方式实现。以上所描述的装置实施例仅仅是示意性的，例如，附图中的流程图和框图显示了根据本申请的多个实施例的装置、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或代码的一部分，所述模块、程序段或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现方式中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。In the embodiments provided in this application, it should be understood that the disclosed apparatus and method may also be implemented in other manners. The apparatus embodiments described above are merely illustrative, for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architectures, functions and possible implementations of apparatuses, methods and computer program products according to various embodiments of the present application. operate. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more functions for implementing the specified logical function(s) executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented in dedicated hardware-based systems that perform the specified functions or actions , or can be implemented in a combination of dedicated hardware and computer instructions.
另外，在本申请各个实施例中的各功能模块可以集成在一起形成一个独立的部分，也可以是各个模块单独存在，也可以两个或两个以上模块集成形成一个独立的部分。In addition, each functional module in each embodiment of the present application may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.
所述功能如果以软件功能模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .
综上所述，本申请实施例提供的一种基于多特征融合的相似性图像检索方法及系统，其能够利用SIFT和GIST特征算子相结合的方式，更加充分地提取了图像的局部和全局特征，并进行了较为充分的融合；利用自适应权重的方式，在相似度计算的过程中，考虑了不同方法。To sum up, the method and system for similarity image retrieval based on multi-feature fusion provided by the embodiments of the present application can more fully extract the local and global image features by combining SIFT and GIST feature operators. features, and have been fully fused; using the adaptive weight method, different methods are considered in the process of similarity calculation.
以上所述仅为本申请的优选实施例而已，并不用于限制本申请，对于本领域的技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。The above descriptions are only preferred embodiments of the present application, and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included within the protection scope of this application.
对于本领域技术人员而言，显然本申请不限于上述示范性实施例的细节，而且在不背离本申请的精神或基本特征的情况下，能够以其它的具体形式实现本申请。因此，无论从哪一点来看，均应将实施例看作是示范性的，而且是非限制性的，本申请的范围由所附权利要求而不是上述说明限定，因此旨在将落在权利要求的等同要件的含义和范围内的所有变化囊括在本申请内。不应将权利要求中的任何附图标记视为限制所涉及的权利要求。It will be apparent to those skilled in the art that the present application is not limited to the details of the above-described exemplary embodiments, but that the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Accordingly, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the application is to be defined by the appended claims rather than the foregoing description, which is therefore intended to fall within the scope of the claims. All changes that come within the meaning and scope of equivalents to are included in this application. Any reference signs in the claims shall not be construed as limiting the involved claim.
Claims (10)
Translated fromChinesePriority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010725870.0ACN111914921A (en) | 2020-07-24 | 2020-07-24 | A method and system for similarity image retrieval based on multi-feature fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010725870.0ACN111914921A (en) | 2020-07-24 | 2020-07-24 | A method and system for similarity image retrieval based on multi-feature fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111914921Atrue CN111914921A (en) | 2020-11-10 |
Family
ID=73280266
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010725870.0APendingCN111914921A (en) | 2020-07-24 | 2020-07-24 | A method and system for similarity image retrieval based on multi-feature fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111914921A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112906696A (en)* | 2021-05-06 | 2021-06-04 | 北京惠朗时代科技有限公司 | English image region identification method and device |
| CN113158745A (en)* | 2021-02-02 | 2021-07-23 | 北京惠朗时代科技有限公司 | Disorder code document picture identification method and system based on multi-feature operator |
| CN113191277A (en)* | 2021-05-06 | 2021-07-30 | 北京惠朗时代科技有限公司 | Table image region identification method and system based on entropy check |
| CN113918744A (en)* | 2021-09-13 | 2022-01-11 | 北京房江湖科技有限公司 | Similar image retrieval method, device, storage medium and computer program product |
| CN115357742A (en)* | 2022-08-02 | 2022-11-18 | 广州市玄武无线科技股份有限公司 | Image plagiarism check method, system, terminal device and storage medium for stores |
| CN116061187A (en)* | 2023-03-07 | 2023-05-05 | 睿尔曼智能科技(江苏)有限公司 | Method for identifying, positioning and grabbing goods on goods shelves by composite robot |
| CN116597176A (en)* | 2023-05-26 | 2023-08-15 | 重庆长安汽车股份有限公司 | Image deduplication method, device, equipment and storage medium |
| CN117354471A (en)* | 2023-12-05 | 2024-01-05 | 深圳市微浦技术有限公司 | Multi-camera collaborative monitoring method, device, equipment and storage medium |
| CN118069870A (en)* | 2024-01-25 | 2024-05-24 | 飞虎互动科技(北京)有限公司 | Method, system and equipment for establishing environment information file based on information fusion characteristics |
| CN118334383A (en)* | 2024-06-13 | 2024-07-12 | 山东智航智能装备有限公司 | Method and device for updating change detection sample |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596277A (en)* | 2018-05-10 | 2018-09-28 | 腾讯科技（深圳）有限公司 | A kind of testing vehicle register identification method, apparatus and storage medium |
| CN110765286A (en)* | 2019-09-09 | 2020-02-07 | 卓尔智联(武汉)研究院有限公司 | Cross-media retrieval method and device, computer equipment and storage medium |
- 2020
- 2020-07-24CNCN202010725870.0Apatent/CN111914921A/enactivePending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596277A (en)* | 2018-05-10 | 2018-09-28 | 腾讯科技（深圳）有限公司 | A kind of testing vehicle register identification method, apparatus and storage medium |
| CN110765286A (en)* | 2019-09-09 | 2020-02-07 | 卓尔智联(武汉)研究院有限公司 | Cross-media retrieval method and device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 薛超等: "基于多核学习的GIST全局和SIFT局部特征融合遥感图像检索方法", 《青岛大学学报（自然科学版）》, vol. 33, no. 1, pages 5 - 11* |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158745A (en)* | 2021-02-02 | 2021-07-23 | 北京惠朗时代科技有限公司 | Disorder code document picture identification method and system based on multi-feature operator |
| CN113158745B (en)* | 2021-02-02 | 2024-04-02 | 北京惠朗时代科技有限公司 | Multi-feature operator-based messy code document picture identification method and system |
| CN113191277B (en)* | 2021-05-06 | 2023-12-19 | 北京惠朗时代科技有限公司 | Table image area identification method and system based on entropy verification |
| CN113191277A (en)* | 2021-05-06 | 2021-07-30 | 北京惠朗时代科技有限公司 | Table image region identification method and system based on entropy check |
| CN112906696A (en)* | 2021-05-06 | 2021-06-04 | 北京惠朗时代科技有限公司 | English image region identification method and device |
| CN113918744A (en)* | 2021-09-13 | 2022-01-11 | 北京房江湖科技有限公司 | Similar image retrieval method, device, storage medium and computer program product |
| CN115357742A (en)* | 2022-08-02 | 2022-11-18 | 广州市玄武无线科技股份有限公司 | Image plagiarism check method, system, terminal device and storage medium for stores |
| CN116061187B (en)* | 2023-03-07 | 2023-06-16 | 睿尔曼智能科技(江苏)有限公司 | Method for identifying, positioning and grabbing goods on goods shelves by composite robot |
| CN116061187A (en)* | 2023-03-07 | 2023-05-05 | 睿尔曼智能科技(江苏)有限公司 | Method for identifying, positioning and grabbing goods on goods shelves by composite robot |
| CN116597176A (en)* | 2023-05-26 | 2023-08-15 | 重庆长安汽车股份有限公司 | Image deduplication method, device, equipment and storage medium |
| CN117354471A (en)* | 2023-12-05 | 2024-01-05 | 深圳市微浦技术有限公司 | Multi-camera collaborative monitoring method, device, equipment and storage medium |
| CN117354471B (en)* | 2023-12-05 | 2024-04-09 | 深圳市微浦技术有限公司 | Multi-camera collaborative monitoring method, device, equipment and storage medium |
| CN118069870A (en)* | 2024-01-25 | 2024-05-24 | 飞虎互动科技(北京)有限公司 | Method, system and equipment for establishing environment information file based on information fusion characteristics |
| CN118334383A (en)* | 2024-06-13 | 2024-07-12 | 山东智航智能装备有限公司 | Method and device for updating change detection sample |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111914921A (en) | A method and system for similarity image retrieval based on multi-feature fusion | |
| US10949702B2 (en) | System and a method for semantic level image retrieval | |
| CN108288088B (en) | A scene text detection method based on end-to-end fully convolutional neural network | |
| CN114897806B (en) | Defect detection method, electronic device and computer readable storage medium | |
| CN111814902A (en) | Target detection model training method, target recognition method, device and medium | |
| CN110175615B (en) | Model training method, domain-adaptive visual position identification method and device | |
| CN111079785A (en) | An image recognition method, device and terminal device | |
| Li et al. | Robust vehicle detection in high-resolution aerial images with imbalanced data | |
| Manivannan et al. | HEp-2 cell classification using multi-resolution local patterns and ensemble SVMs | |
| KR20130142191A (en) | Robust feature matching for visual search | |
| CN114595352B (en) | Image recognition method and device, electronic equipment and readable storage medium | |
| CN107633226A (en) | A kind of human action Tracking Recognition method and system | |
| CN104751475B (en) | A kind of characteristic point Optimum Matching method towards still image Object identifying | |
| CN110532413B (en) | Information retrieval method and device based on picture matching and computer equipment | |
| CN115375917B (en) | Target edge feature extraction method, device, terminal and storage medium | |
| CN106295498A (en) | Remote sensing image target area detection apparatus and method | |
| CN110852327A (en) | Image processing method, device, electronic device and storage medium | |
| CN115082781A (en) | Ship image detection method and device and storage medium | |
| CN111815582A (en) | A two-dimensional code region detection method with improved background prior and foreground prior | |
| CN115690803A (en) | Digital image recognition method, device, electronic device and readable storage medium | |
| CN111738319B (en) | Clustering result evaluation method and device based on large-scale samples | |
| CN115564778B (en) | Defect detection method and device, electronic equipment and computer readable storage medium | |
| Wang et al. | Superpixel-level target discrimination for high-resolution SAR images in complex scenes | |
| CN110135428B (en) | Image segmentation processing method and device | |
| CN112214639A (en) | Video screening method, video screening device and terminal equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20201110 |