Research On Emotion Classification Of Yunnan Folk Oil Painting

Painting is the carrier of the artist’s artistic emotions.Therefore,the emotional classification of painting images is an important aspect in the management of painting images.The research on painting emotion classification has become a research hot-spot in computer visual information processing.Yunnan oil painting is an important part of Chinese painting art.It always concern folklore and landscape,fully demonstrating the rich and colorful regional culture of Yunnan.The existing painting sentiment classification algorithm has serious shortage of face area features,and the classification algorithm is inefficient,which is difficult to directly apply to the emotional classification of Yunnan oil painting.This thesis takes the emotional classification of Yunnan folk oil painting as the research object,and focuses on the application of related image feature extraction and sentiment classification in Yunnan folk oil painting.The algorithm of emotional classification of Yunnan folk oil painting based on extreme learning machine is proposed.The specific content of this thesis is as follows:Firstly,aiming at the difficult problems of face feature extraction,this thesis proposes a face extraction algorithm based on face localization and wavelet transform for Yunnan oil painting.The algorithm firstly uses the rectangular feature to comprehensively detect the face in the oil painting,and then combines the binarization and gray projection method to locate the key parts of the face.On this basis,the face feature information is extracted by Gabor wavelet transform,and the face feature information in Yunnan oil painting image is reduced based on PCA principal component analysis.Secondly,an algorithm for emotional classification of Yunnan folk oil painting based on Exreme Learning Machine is proposed.Compared with the traditional SVM or BP neural network based sentiment classification algorithm,the training speed is slow and easy to fall into the local optimal solution.ELM takes into account the dual requirements of computer training rate and classification accuracy,and has more practicability.The experimental results show that this thesis studies the emotional classification algorithm of Yunnan folk oil painting based on ELM,which simplifies the algorithm flow and improves the accuracy of sentiment classification.Finally,the design and implementation of the sentiment classification system.The system uses Matlab language combined with its human-computer interaction interface(GUI)design function to develop,and then uses the painting images in the self-built database for experimental testing,and compared with other methods based on the recognition rate performance.The experimental results show that the extraction of the features of the face region in oil painting can effectively improve the accuracy of oil painting emotion classification.