Image feature extraction device, image feature extraction method, and image recognition device

An image feature extraction method comprising:

• generating intensity gradient data with respect to two different directions based on intensity data of an image data;
• calculating a covariance matrix for each partial area obtained by dividing the image data based on the intensity gradient data with respect to the two different directions;
• calculating two parameters related to a major axis and a minor axis of an ellipse expressed by the covariance matrix;
• quantizing a range of the logarithms of the parameters for each of the partial area using a predetermined division number;
• and
• outputting a feature vector which contains a value only at a dimension corresponding to the quantized range different from the other dimensions.

An image recognition device comprising:

• an image data input unit acquires image data and outputs intensity data based on the image data;
• a gradient image calculator generates intensity gradient data with respect to two different directions based on the intensity data;
• a gradient count unit calculates a covariance matrix for each partial area obtained by dividing the image data into a plurality of areas based on the intensity gradient data;
• a feature data output unit calculates two parameters related to a major axis and a minor axis of an ellipse expressed by the covariance matrix, quantizes a range of the logarithms of the parameters for each of the partial area using a predetermined division number, and outputs a feature vector which contains a value only at a dimension corresponding to the quantized range different from the other dimensions;
• and
• an image recognition unit identifies an image based on the feature vectors.