Papers by Keyword: Skeletonization

Abstract: The main objective of this study is to discover and investigate greater levels of human motion activities recognition. The study presents four approaches of human motion data processing to recognize the human activities. Data collection process was performed in two ways: wearable sensor based in signal data and vision based in image data. The proposed approaches used to analyze the signal and image data are: wearable sensor using 3-space sensing with angular velocity and elevation angle as moderators, wearable sensor using statistical nine existing and a proposed developed classifiers as classification learning system, vision based using skeletonization with humerus-radius and horizontal-radius as measuring angle and vision based image-signal histogram using 2D-1D transformation method. The principal contributions of this thesis are the development of the human motion analysis methods with validated evaluation process tested on the proposed systems. The proposed systems achieved more than 98 % for signal processing and 97 % for image processing of accuracy on recognizing human activities.

Abstract: As One of Features from other Languages, the Chinese Tone Changes of Chinese are Mainly Decided by its Vowels, so the Vowel Variation of Chinese Tone Becomes Important in Speech Recognition Research. the Normal Tone Recognition Ways are Always Based on Fundamental Frequency of Signal, which can Not Keep Integrity of Tone Signal. we Bring Forward to a Mathematical Morphological Processing of Spectrograms for the Tone of Chinese Vowels. Firstly, we will have Pretreatment to Recording Good Tone Signal by Using Cooledit Pro Software, and Converted into Spectrograms; Secondly, we will do Smooth and the Normalized Pretreatment to Spectrograms by Mathematical Morphological Processing; Finally, we get Whole Direction Angle Statistics of Tone Signal by Skeletonization way. the Neural Networks Stimulation Shows that the Speech Emotion Recognition Rate can Reach 92.50%.

Abstract: This paper presents an image based numerical method proposed to obtain information regarding pore structure and organization of pores within materials based on 3D digital image input. The output of the numerical algorithm is a pore size distribution of materials. The algorithm is based on the combination of the two digital image processing algorithms: 1) a medial axis thinning algorithm to obtain 3D skeleton of the pore structure, and 2) the distance transform of an image. The method is tested on simple 2D and 3D microstructures of packed spheres, demonstrating the performance of the proposed method.

Abstract: Vectorization is the most fundamental operation in interpretation of line drawings and document analysis. There are several reasons for converting image vectorization. Vector data is normally created from existing natural source image like photographs,scanned images. Choosing a best vectorization method that suits the needs of the system is very important. In general, good methods must preserve information like line geometry and intersection junction as far as possible. It is also important to analyze the error and find the accuracy of the result with respect to the original data. We have compared Skeletonization by Mathematical Morphology and Voronoi Diagrams with original image for vectorizing images. Root mean squre error is one of the good methods to analysis an error on original Image, Mathematical Morphology and Voronoi Diagrams. Literature about above said methods is also included in this paper.

Abstract: Image processing toolbox of Matlab provides a morphological skeletonization operator. But this function is sensitive to noises. The aliasing shape's skeleton computed by this operator involves many extra branches. In order to eliminate the influence of aliasing noises, a combined method is introduced. The method combines the morphological thinning and skeletonization operators, and takes their advantages. It is not sensitive to noises, and can eliminate the extra branches.