Authors: Xiaofeng Liu
Addresses: Academy of Fine Arts and Design, Mudanjiang Normal University, Mudanjiang 157011, China
Abstract: Aiming at the problem of time-consuming and large error in current image feature extraction results, a symbolic feature extraction method for art graphics based on time constraint is proposed. Gauss filtering method is used to filter out the image noise, and Laplacian sharpening operation is used to enhance the edge of graphic symbols. According to the enhancement results, an active contour model is introduced to define a closed contour, and a driving force is used to realise the evolution of the contour to the target boundary to complete the graph segmentation. The geometric transformation of the segmentation results is analysed by translation transformation, scale transformation and rotation transformation. Based on graphics enhancement, segmentation and geometric transformation analysis, time constraints are introduced. By analysing the mean value and variance level of time-constrained laminar flow features, the average value and divergence degree of time-constrained laminar flow features are analysed, and the time-constrained symbols are regarded as a vector to realise the symbolic feature extraction of fine arts graphics. The experimental results show that the proposed method is efficient and accurate, and it is a reliable method for extracting symbolic features of fine arts graphics.
Keywords: image noise; time constraint; art graphics; symbolisation; feature extraction; geometric transformation.
International Journal of Information and Communication Technology, 2020 Vol.17 No.2, pp.194 - 209
Received: 12 Jun 2019
Accepted: 25 Jul 2019
Published online: 09 Jul 2020 *