Osaka Prefecture University Public Corporation

Olympus Corporation

Masakazu Iwamura

Koichi Kise

Kazuto Noguchi

Yukihito Furuhashi

Taiji Mine

Tomohiro Nakai

An image retrieval method comprising: a step of extracting at least one query feature vector from a query image on which a subject of the image retrieval is captured, the query feature vector representing a local feature of the query image; a step of accessing an image data base in which a plurality of reference images are stored previously, each reference image being stored in conjunction with learning images generated therefrom and reference feature vectors representing local features of the reference image and the learning images; a comparing step of comparing the query feature vector with the reference feature vectors stored in conjunction with each reference image using an approximate nearest neighbor search to find a reference feature vector approximately nearest to the query feature vector; and a selecting step of selecting a reference image with which the found reference feature vector is stored in conjunction from the reference images as a retrieval result wherein: the learning image is generated by adding a defocus and/or a motion-blur effect likely to occur on capturing the subject to each reference image, the reference feature vectors are extracted from each reference image and the learning image corresponding to the reference image respectively using the scale-space approach, the query feature vector is extracted from the query image using the scale-space approach, and each of the above steps is executed by a computer.

Image retrieval method, image retrieval program, and image registration method

A pattern recognition apparatus including: an extracting section for extracting, from a query image that is composed of at least one piece of pattern component and previously undergoes a geometric transformation, the pattern component; a feature acquiring section for acquiring a geometric invariant feature of the pattern component as a query feature, the query feature being represented by at least three feature points including first, second and third feature points, each feature point locating on the pattern component and being retrieved from the pattern component based on a predetermined rule; a comparing section for comparing the query feature with a plurality of reference features, each reference feature representing different reference patterns prepared as candidates for pattern recognition; and a pattern determination section for determining, as a recognition result, a specific reference pattern out of the candidates based on a similarity of features therebetween and, wherein: each reference feature is represented using feature points retrieved from each reference pattern based on the same rule as that of the query feature, and based on the predetermined rule, a position of the first feature point is specified out of points which locate on the pattern component and are invariant to the geometric transformation, a position of the second feature point is specified using a characteristic regarding a shape of the pattern component, the characteristic being invariant to the geometric transformation, and a position of the third feature point is specified from a predetermined value being invariant to the geometric transformation and from the specified positions of the first and second feature points.

Pattern recognition apparatus

Provided is a method for constructing an image database for object recognition, which includes a feature extraction step of extracting local descriptors from object images which are to be stored in an image database, a scalar quantization step of quantizing a numeric value indicating each dimension of each of the local descriptors into a predetermined number of bit digits, and a storing step of organizing each of the local descriptors after the quantization to be able to be searched for in the closest vicinity, giving to the local descriptor an identifier of the image from which the local descriptor has been extracted, and storing the local descriptor to which the identifiers are given in the image database. The storing step comprises extracting the local descriptors from the object images when a search query is given, scalar-quantizing each dimension, determining a local descriptor in the closest vicinity of each of the local descriptors from the image database, and storing each local descriptors so as to be able to identify one image by majority vote processing from the images including any determined local descriptor. The scalar quantization step comprises quantizing each dimension of each of the local descriptors into 8 bits or less. Also provided are a processing program for the method and a processing device for performing the processing.

Method for constructing image database for object recognition, processing apparatus and processing program

For object recognition based on nearest neighbor search of local descriptors such as SIFT, it is important to keep the nearest neighbor search efficient to deal with a huge number of descriptors. The present invention provides methods of efficient recognition. In one embodiment, the method is based on the observation that the level of accuracy of nearest neighbor search for correct recognition depends on images to be recognized. The method is characterized by the mechanism that multiple recognizers with approximate nearest neighbor search are cascaded in the order of the level of approximation so as to improve the efficiency by adaptively controlling the level to be applied depending on images. In another embodiment the method is characterized by excluding local descriptors with low discriminability when a plenty of local descriptors are present in the vicinity and a plenty of distance calculation are required.

Image recognition method, image recognition device, and image recognition program

A document/image retrieval method for retrieving a document/image corresponding to a captured digital image from a database by comparing features calculated based on feature points of the captured digital image with features preliminarily calculated based on feature points of each of documents and/or images stored in the database, the method comprising: extracting the feature points from the captured digital image; defining a local set of feature points for each of the extracted feature points; selecting feature points from the defined local set to define a feature point subset of the local set; determining invariant values as values characterizing the defined subset for combinations of the feature points in the subset, the invariant values being invariant to a geometric transformation; calculating a feature by combining the determined invariant values; and performing a voting process on the documents and/or images in the database based on the preliminarily calculated features of the documents and/or images; whereby the document/image corresponding to the captured digital image is retrieved from the database.

Method, program and apparatus for storing document and/or image using invariant values calculated from feature points and method, program and apparatus for retrieving document based on stored document and/or image

Growing community of inventors

Sakai, Japan

Masakazu Iwamura