Directionality Of 3D Surface Textures By Using Adaptive Directional Lifting Wavelet Tranform

Directionality measurement of texture has become an active research area in recent years and is playing a significant role in numerous fields of image processing, image retrieval, computer vision, pattern recognition, and computer graphics, which demonstrates a great prospect of application of it. Traditional research achievements of texture directionality including the intrinsic directionality of texture and illuminate direction are mainly restricted in 2D surface textures, and little research has been done on 3D surface texture.In fact, real-world textures are 3D surface textures. Research on 3D surface texture contains appearance representation under different illumination and view conditions, data collection, analysis synthesis and visualization. In this thesis, directionality of 3D surface textures is measured by pointed method, and the measured results are analyzed and evaluated.This thesis first introduces texture, especially 3D texture, and presents current situation of texture directionality research. Then the background knowledge and relevant research methods are stated, and the 3D surface texture directionality research is further discussed. Making feature extraction by adaptive directional lifting wavelet and being based on some measurement algorithms of 2D surface textures directionality research for reference, the thesis applies adaptive directional lifting wavelet transform to the experiments of 3D surface texture directionality detection. According to current various methods on texture classification, the classification algorithm of 3D surface texture based on adaptive directional lifting wavelet transform is proposed. By comparing the test set and train set, the 3D surface texture is classified by adopting different similarity measurement algorithms. Having achieved ideal results, the experiments provide future research work with new thoughts and methods. On same texture's the different texture appearances under different illumination directions, the thesis utilizes adaptive directional lifting wavelet transform to extract features of 3D surface texture, classifies the 3D surface texture by illumination, and detect the illumination direction of 3D surface texture. Further more, given that lifting wavelet can be constructed perfectly, it is applied to the edit of 3D surface texture and new images different from the original texture features are generated. Finally, the research work of 3D surface texture direction by using adaptive directional lifting wavelet transform is summed up; the innovation and deficiency of the thesis and the prospect of future work are generalized.