Dissertation > Excellent graduate degree dissertation topics show
Study on Natural Texture Image Synthesis
Author: LiWenXiu
Tutor: BiXiaoJun
School: Harbin Engineering University
Course: Signal and Information Systems
Keywords: Texture characteristic function BP Network Image generation
CLC: TP391.41
Type: Master's thesis
Year: 2005
Downloads: 113
Quote: 2
Read: Download Dissertation
Abstract
|
Texture image generation technology is one of the important topics of the research field of digital image processing , computer graphics , it can be widely used in areas such as communications , security , and has a high value . Texture generation technology is not yet mature , existing methods can not meet the actual requirements to generate a variety of texture image . The main contents of the subject is to be applied to theoretical approaches in the field of artificial intelligence is more mature natural texture image generator , natural texture images generated using artificial neural network algorithm . The artificial neural network BP algorithm can effectively solve the non-linear mathematical modeling problem , and has a strong parallel processing capabilities . The topics include the following three aspects : First, the characteristic function of the natural texture images in-depth research and analysis , and to determine from all angles, true fully reflect the characteristics of the image , and finally selected 16 characteristic function as a basis for image generation ; second, to establish effective on BP network model , 16 of the original image characteristic function as the input of BP network , training by learning to identify artificial neural network structure ; third , the formation of a natural texture image generation software . When given the natural texture images , the software will automatically extraction of the characteristic function , automatically generate the image of the artificial neural network architecture , and finally achieve the image generated by learning - training . Experimental results show that the texture image generation method proposed by the present subject having a short generation time , the generation effect lifelike . The method to break the boundaries of the texture image based on different structures with different pattern generation method , freely generate a variety of natural texture images .
|
Related Dissertations
- BP network optimization based on genetic algorithm optimization of the biodiesel process,TE667
- DSP deep measurements tracked robot obstacle avoidance system,TP242
- Research and Development of Product Oil Depot Automation Supervisory and Control System,TP273.5
- Research on Carbonation of Concrete Materials in Qinghai’s Salt Lake,TU528
- Research and Realization of the Product Styling Design CAD System Based on Product Image,TB472
- The Study of Toll Audit in Networked Expressway Based on Neural Network,U495
- Steam Generator Fault prediction,TL353.13
- Research on Microscopic Evaluation Method and Application of Freeway Traffic Safety,U492.8
- Hybrid optimization algorithm for parameter optimization of an injection mold,TP391.72
- Research of Aero-engine Fault Diagnosis Based on Rough Neural Network,V263.6
- Research on Anti-Cracking Performances of the Cement Stable Macadam Adding Admixture,U414
- Digital image forensics technology research,TP391.41
- Applied Research of Support Vector Regression Machine Based on Parameter Optimization,TH823
- VQ-based model and BP network of high naturalness speech,TN912.3
- Study of Market Demand Prediction System on ANN in Auto Parts,TP311.52
- Based on Neural Network tooth root stress calculation method,TG711
- Modeling Analysis of Thermal Characteristic of Solar Collectors Based on Artificial Neural Network,TK513
- Genetic algorithms and neural network stock prediction analysis,F224
- Metering Heating System Operation Adjustment Were Discussed,TU832
- Research and Application of an Improved Genetic Algorithm Optimizated BP Network,TP183
- The model of credit rank evaluation of medical organizations appointed by medicare center based on Artificial Neural Networks,TP183
CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net Mobile
|