SURF feature extraction algorithm based on visual saliency improvement

Zongyuan Zhu, Guicang Zhang, Hongjie Li

Feature extraction is an important link in image retrieval and image matching.Aiming at the problem of the traditional feature extraction method, which is too simple to extract valid dimension and feature points, a SURF weight algorithm combining visual significance and improvement is proposed to extract the key points:. SURF algorithm of image is utilized to extract the key point, and then through improved significant find significant area detection method, key points can be divided into two parts, the significant region and significant area, external point by weighting algorithm to judge the importance of structure information, thus retaining structure information important points.The experimental results show that the feature points extracted by this method are more comprehensive and improve the accuracy of image matching.