Method and system for sample image index creation and image filtering and search

A method for image filtering, the method comprising:

• under control of one or more servers comprising one or more processors;
• establishing a sample image index table, the sample image index table storing a correlation between node identifications of a plurality of nodes, an identification of a sample image having regional characteristics that are clustered into the plurality of nodes, and a number of times that respective regional characteristics of each of the plurality of nodes appear in the sample image, the sample image index table further storing node identifications of the nodes of the sample image and a count of a number of the nodes in the sample image;
• extracting regional characteristics from an image to be searched;
• clustering the regional characteristics of the image to be searched into corresponding nodes;
• obtaining a corresponding sample image identification by indexing the sample image index table using node identifications of the nodes of the image to be searched;
• determining a number of duplicate nodes between the image to be searched and the sample image;
• comparing a first number of nodes of the images to be searched and a second number of nodes of the sample image to determine the first number or the second number;
• obtaining a degree of similarity degree between the image to be searched and the sample image as a ratio of the number of duplicate nodes between the image to be searched and the sample image to the first number if the first number is determined or the second number if the second number is determined;
• and
• filtering out the image to be searched in an event that a degree of similarity between the image to be searched and the sample image exceeds a similarity threshold.

An apparatus for creating a sample image index table, the apparatus comprising:

• a characteristics extraction unit that extracts regional characteristics from a sample image by using an extraction algorithm with scale invariability;
• a clustering unit that clusters the regional characteristics of the sample image into a plurality of nodes each of which having a respective number of regional characteristics having a same degree of similarity, each node corresponding to a respective node identification;
• and
• a storage unit that establishes the sample image index table, the sample image index table storing a correlation between a first node identification of a first node of the plurality of nodes, an identification of the sample image, and a number of times that the respective regional characteristics of the first node appear in the sample image, the sample image index table further storing node identifications of the nodes of the sample image and a count of a number of the nodes in the sample image.

An apparatus for image filtering, the apparatus comprising:

• a storage unit that stores a sample image index table, the sample image index table storing a correlation between node identifications of a plurality of nodes, an identification of a sample image having regional characteristics that are clustered into the plurality of nodes, and a number of times that respective regional characteristics of each of the nodes appear in the sample image, the sample image index table further storing node identifications of the nodes of the sample image and a count of a number of the nodes in the sample image;
• an extraction unit that extracts regional characteristics from an image to be searched;
• a node determination unit that clusters the regional characteristics of the image to be searched into corresponding nodes;
• an index unit that obtains a corresponding sample image identification by indexing the sample image index table using node identifications of the nodes of the image to be searched;
• a calculation unit that calculates a number of duplicate nodes between the image to be searched and the sample image;
• a similarity degree calculation unit that:
• compares a first number of nodes of the images to be searched and a second number of nodes of the sample image to determine the first number or the second number, and
• obtains the degree of similarity degree between the image to be searched and the sample image as a ratio of the number of duplicate nodes between the image to be searched and the sample image to the first number if the first number is determined or the second number if the second number is determined;
• and
• an output filtering unit that filters out the image to be searched in an event that the degree of similarity between the image to be searched and the sample image exceeds a similarity threshold.

A method for image searching, the method comprising:

• under control of one or more servers comprising one or more processors;
• establishing a sample image index table, the sample image index table storing a correlation between node identifications of a plurality of nodes, an identification of a sample image having regional characteristics that are clustered into the plurality of nodes, and a number of times that respective regional characteristics of each of the nodes appear in the sample image, the sample image index table further storing node identifications of the nodes of the sample image and a count of a number of the nodes in the sample image;
• extracting regional characteristics from an image to be searched;
• clustering the regional characteristics of the image to be searched into corresponding nodes;
• obtaining a corresponding sample image identification by indexing the sample image index table using node identifications of the nodes of the image to be searched;
• calculating a number of duplicate nodes between the image to be searched and the sample image;
• comparing a first number of nodes of the images to be searched and a second number of nodes of the sample image to determine the first number or the second number;
• obtaining a degree of similarity degree between the image to be searched and the sample image as a ratio of the number of duplicate nodes between the image to be searched and the sample image to the first number if the first number is determined or the second number if the second number is determined;
• and
• providing the sample image as a search result in an event that the degree of similarity between the image to be searched and the sample image exceeds a similarity threshold.