The 52nd. IEEE International Midwest Symposium on Circuits and Systems (MWSCAS 2009) - Cancún, México
Analog Design for Human-Centric Smart Living Technologies
Prof. Franco Maloberti, University of Pavia, Italy.
Abstract: In the fast growing electronic world, analog integrated circuits are playing an important role for sensing devices and networking, security and safety, healthcare medical and life science, entertainments and education, and many other activities that improve the quality of human life. The use of digital methodologies and powerful digital processor is the core of modern electronic systems but the analog nature of the real world makes key the analog interfaces and analog operations.
Many portable or nomadic electronic systems will need to acquire power from the environment without the need of refueling. Energy can come from different and multiple sources but in addition to suitable devices for power harvesting it is necessary to design effective electronics that transform and conveniently store energy. The new solutions will be able to obtain high efficiency with limited use of bulky devices like inductors and capacitors. The optimal solution will involve technological and economical aspects and, among them the choice between system on chip (SoC) or system in package (SiP), choice that will be driven by cost effectiveness, required complexity, time-to-market and miniaturization requests.
Vector Order Statistics and Fuzzy Set Approaches in 3D Digital Processing of  Multichannel Images and Video Sequences Employing DSP/FPGA
Dr. Volodymyr Ponomaryov, National Polytechnic Institute of Mexico, ESIME-Culhuacan
Abstract: Satellite, Radar, Medical, Digital Photographs, High Definition Television, Virtual Reality, Electron Microscopy, etc, are some of the multispectral and multichannel image applications that need application of restoration and denoising procedures. All these images usually are corrupted by noise due to sensors influence, during transmission of the signals, or noise may be produced by environmental phenomena too (e.g. atmospheric noise), or can be affected by multiplicative noise (speckle). So, the application of efficient algorithms is necessary to decrease influence produced by a noise that limit quality of an image. The main objective of present paper to expose the justified novel approach in restoration of the multichannel and multispectral images that can be used in mentioned applications.
      So, it is necessary to have several efficient denoising schemes, which depend on noise type and other priory information, in a pre-processing stage of a general image processing system. The effectiveness of different restoration algorithms depends on a technique used where the existed powerful approaches have been probed in processing the “true color†images but not in multispectral applications. Here, we present novel approach that generates several frameworks in processing the multispectral images adapting it to multichannel and multispectral bands, and finally, justifying them in numerous simulations where they demonstrated excellent performance properties. The idea employed is consists of usage of two most powerful techniques; those are: the fuzzy set theory and directional techniques that follow from vector order statistics theory. By now, these two techniques have been applied in numerous filtering algorithms separately. Moreover, they present good results outperforming other techniques in different noise scenarios. Unfortunately, a methodology, which gathers the advantages of each one of these techniques, that provides the better suppression noise capability, fine details preservation as well as chromaticity characteristics, is not exist. So, this research presents and justifies a new methodology based on connection of both techniques to realize enhancing the rendering for each filter used in independent way.
     General operations of novel approach consist of in estimation made in fuzzy means for any spectral band: if there are existed the edges and fine details, or impulsive noise, or may be some movement in the central pixel into sliding window used in processing. So, it is possible to distinguish these characteristics inherent in multispectral images using fuzzy rules designed and presented in this work. They are applied to fuzzy-directional values to give the certainly: if the central pixel component is a corrupted one or not. In case of a corrupted pixel happened, some procedures in substitution of the central component with one of its neighbours, based on special it selection, are justified using found fuzzy rules values, which indicate the level of pixel corruption.Â
     Different test and real mages and sequences, which have been used, show the effectiveness of proposed approach. Mentioned visual data include: commonly employed test color images: “Lenaâ€, “Baboonâ€, “Peppersâ€, “Goldhillâ€, and “Parrotsâ€, etc.; satellite multispectral real images for the same scene, which were received in several Landsat 7 satellite spectral bands (channels); and, finally known video sequences: “Miss Americaâ€, “Flowers†and “Chairâ€, etc. Experimental justification of the proposed technique in numerous video image data with different texture characteristics permits a better understanding the robustness of the novel frameworks.
Several designed promising algorithms as well as better existed ones were implemented on the DSP/FPGA platforms realizing analysis of the sequences or images in a real time environment.
Matthew T. Hunter, Ph.D. Chief Research Scientist Signal Processing and Communications DME Corporation
Check out FPGA related videos
Find Us On Facebook
FPGA Central is a website bringing the FPGA (Field Programable Gate Array), CPLD , PLD, VLSI community together at one central location. Some Keywords: FGPA, FPGA, EDA Tools, FPGA Design, Central, Programmable logic, LUT, VLSI, SoC, Journal
Privacy Policy || Terms of Use
 |
 |