Computer device and method for training feature point detector and feature point detection method

Abstract

The invention discloses a computer device and method for training a feature point detector and an image feature point detection method. The computer device comprises an image transformation module, afeature point detector and a model training module, the image transformation module is used for carrying out random sampling transformation on an original image based on an image transformation operation set to obtain a transformed image, and the model training module is used for carrying out global batch normalization operation on each differential response graph and training each differential response graph by utilizing an unsupervised machine learning method; and the feature point detector is used for processing the input image by using two convolution kernels with different scales to obtain respective corresponding response feature maps, and carrying out differential calculation on the output response feature maps of the convolution kernels to obtain a differential response map of theoriginal image and the transformed image. According to the technical scheme, the data processing flow is simple, the invention is more sensitive to the edge of the object, the feature point dense selection phenomenon can be effectively relieved, and the method is suitable for small-scale data set application scenes.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, in particular to a computer device and method for training a feature point detector and an image feature point detection method. Background technique [0002] With the rapid development of computer vision technology, image processing technology is also facing severe challenges. Image feature points contain rich local image features and provide helpful special points for subsequent image processing tasks, such as object edge points, corner points, textures, spots, etc. Excellent image features can help solve computer vision problems such as object recognition, image matching, visual tracking, and 3D reconstruction. Therefore, the feature point detection process is a basic but critical part of computer vision technology. [0003] The traditional feature point detection methods are all based on the artificially designed detector method, and a detector that can detect pixels that m...

AI Technical Summary

Problems solved by technology

However, the training mode of randomly sampling image blocks requires steps such as random selection, positioning, and cropping of image blocks, and the data processing is cumbersome, and the network architecture is complex; in addition, the simple linear filter is easily affected by the scale, and the output response value of the edge part of the object is too large. close to each other, not only makes the positioning accuracy of the real edge pixels poor, but even causes more serious feature point aggregation.

Images