Browsing by Author "Lin, CF"

Now showing 1 - 20 of 66

A high-performance and memory-efficient pipeline architecture for the 5/3 and 9/7 discrete wavelet transform of JPEG2000 codec
(2005-12-01)
Wu, BF
;
Lin, CF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we propose a high-performance and memory-efficient pipeline architecture which performs the one-level two-dimensional (2-D) discrete wavelet transform (DWT) in the 513 and 9/7 filters. In general, the internal memory size of 2-D architecture highly depends on the pipeline registers of one-dimensional (I-D) DWT. Based on the lifting-based DWT algorithm, the primitive data path is modified and an efficient pipeline architecture is derived to shorten the data path. Accordingly, under the same arithmetic resources, the 1-D DWT pipeline architecture can operate at a higher processing speed (up to 200 MHz in 0.25-mu m technology) than other pipelined architectures with direct implementation. The proposed 2-D DWT architecture is composed of two 1-D processors (column and row processors). Based on the modified algorithm, the row processor can partially execute each row-wise transform with only two column-processed data. Thus, the pipeline registers of 1-D architecture do not fully turn into the internal memory of 2-D DWT. For an N x M image, only 3.5N internal memory is required for the 513 filter, and 5.5N is required for the 9/7 filter to perform the one-level 2-D DWT decomposition with the critical path of one multiplier delay (i.e., N and M indicate the height and width of an image). The pipeline data path is regular and practicable. Finally, the proposed architecture implements the 513 and 9/7 filters by cascading the three key components.
A low memory QCB-based DWT for JPEG2000 coprocessor supporting large tile size
(2005)
Wu, BF
;
Lin, CF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
JPEG 2000, which provides a higher compression ratio than the traditional JPEG, is an upcoming compression standard for still images. The experimental results imply that larger tile size used for JPEG 2000 results in better image quality. However, processing the large tile image requires more memory in the hardware implementation. To reduce the hardware resources, a QCB (quad code-block)-based DWT method is proposed to support the process of large tile image with low memory. Based on the QCB-based DWT engine, three code-blocks belonging to LH(0), HL(0) and HH(0) bands can be generated after each fixed time slice recursively and the EBC (embedded block coding) processors can directly process the three code-blocks. It can save the size of tile memory up to 75%. Moreover, the remaining 1/4 size of tile memory can be decreased through the zero holding extension for the unavailable data. That is, it only requires 24K bytes memory to support the process of 512x512 tile image, with slight image degradation, especially at low bit rates. The low memory requirement makes the hardware implementation practicable.
A modified multi-chemical spray cleaning process for post shallow trench isolation chemical mechanical polishing cleaning application
(1998-11-02)
Wang, YL
;
Wang, TC
;
Wu, J
;
Tseng, WT
;
Lin, CF
Chemical mechanical planarization (CMP) has become widely accepted for the formation of device interconnect structures. Shallow trench isolation technology (STI) utilizes CMP and has been applied to deep-sub-micron processes. Poly Si, CVD Si or SiO2 can be grown or deposited in the trench and planarized by a CMP process. However, the typical wafer surface is contaminated with slurry particles and metallic impurities after the CMP process. The silica particles may damage the VLSI patterns and the metallic impurities can induce many crystal defects in Si wafers during the following furnace processing. Therefore, the post CMP clean is a very important step for the STI process. However, the wafer for poly-Si surface is hydrophobic, SiO2 surface is hydrophilic and Si film is very easy to charge up. Thus, the defect can be difficult to remove by a conventional cleaning technique. In this study: we propose the use of a modified multi-chemicals spray cleaning process for post STI CMP cleaning. We used a modified and heated ammonia/peroxide mix (APM) clean with an ammonia pre-soak and an HF step to etch a thin layer for the removal of trapped metallic ions which can be followed by a hydrochloric/peroxide mix (HPM) clean process to assist in the removal of metallic ions. (C) 1998 Elsevier Science S.A. All rights reserved.
A rescheduling and fast pipeline VLSI architecture for lifting-based discrete wavelet transform
(2003)
Wu, BF
;
Lin, CF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we propose a fast pipeline VLSI architecture for ID lifting-based discrete wavelet transform (DWT). This design method merges the filtering steps called the predictor and updater into one single step. Based on this modified algorithm, we explore the data dependency of the input and output signals, and thus make the pipeline design more efficiently for hardware implementation under the same processor elements proposed in previous works. Moreover, the inverse DWT case also adopts the same architecture as the forward DWT. Finally, the area and the working frequency of the proposed architecture in 0.35um technology are 2.511 x 2.510 mm(2), and 150 MHz, respectively.
A study of subbands in AlGaN/GaN high-electron-mobility transistor structures using low-temperature photoluminescence spectroscopy
(2002-06-17)
Fang, CY
;
Lin, CF
;
Chang, EY
;
Feng, MS
;
材料科學與工程學系
Al0.15Ga0.85N/GaN high-electron-mobility transistor (HEMT) structures with various delta-doping concentrations and spacer thicknesses grown on sapphire by metalorganic chemical-vapor deposition are investigated. The Hall mobility is as high as 1333 cm(2)/V s at room temperature and 6330 cm(2)/V s at 77 K. Two-dimensional electron gas (2DEG) phenomena, which have not been clearly resolved in the literature, are observed by photoluminescence (PL) spectra at low temperature in this study. The PL spectra peaks of the interband transitions from 2DEG subbands to the valence band are in the range from 3.486 to 3.312 eV. The effects of the strain caused by different Al fractions of the top layer, and that of the spacer thickness on the 2DEG phenomena are discussed. Redshifts due to temperature variations for various HEMT structures are observed in 2DEG subbands and in the band-edge emission, which is believed to be evidence of interband transitions from 2DEG subbands to valence bands. (C) 2002 American Institute of Physics.
A ULSI shallow trench isolation process through the integration of multilayered dielectric process and chemical-mechanical planarization
(1999-06-22)
Lin, CF
;
Tseng, WT
;
Feng, MS
;
Wang, YL
;
材料科學與工程學系
A multilayer thin film structure (SiO2/SiNx/SiO2) with a one-step chemical-mechanical polish (CMP) process is developed for shallow trench filling and planarization for ULSI devices in the quarter micron regime. By fine-tuning the plasma-enhanced chemical vapor deposition (PECVD) conditions we successfully modified the stoichiometry and other characteristics of the as-deposited SiNx and oxide films. As these film characteristics are changed, the CMP removal rate selectivity between sacrificial oxide and nitride stopper layer can also be adjusted. Correspondingly, the CMP process latitude for shallow trench isolation in 0.25 mu m memory devices can be broadened by sequentially depositing multilayered oxide and nitride films with adjustable characteristics and CMP removal rates. Dishing-free wide trench areas with excellent planarity can be achieved through the integration of the proposed multilayered PECVD oxide/nitride scheme and the one-step CMP process. (C) 1999 Elsevier Science S.A. All rights reserved.
AIGaN/GaN HEMT sub-bands study using low-temperature photoluminescence
(2001)
Fang, CY
;
Lin, CF
;
Lee, CS
;
Chang, EY
;
Feng, MS
Al0.1Ga0.9N/GaN HEMTs (High Electron Mobility Transistor) structures with different delta-doping concentration and spacer thickness grown on sapphire by MOCVD are studied. The hall mobility is as high as 1333 cm(2)/V s. at room temperature and 6330 cm(2)/V s at 77 K. Obvious 2-DEG phenomena not clearly resolved in literatures before are observed by PL spectra measured at low temperature. The influences of delta-doping concentration and spacer thickness on the 2-DEG phenomena are discussed. Peaks red-shift of different HEMT structures due to temperature variation are observed in 2-DEG sub-bands as well as at the band-edge emission that is believed as a strong evidence of 2DEG sub-bands to valence bands transmission.
An efficient architecture for JPEG2000 coprocessor
(2004-11-01)
Wu, BF
;
Lin, CF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
JPEG2000 is a new international standard for still image compression. It provides various functions in one single coding stream and the better compression quality than the traditional JPEG, especially in the high compression ratio. However, the heavy computation and large internal memory requirement still restrict the consumer electronics applications. In this paper, we propose a QCB (quad code block)-based DWT method to achieve the higher parallelism than the traditional DWT approach of JPEG2000 coding process. Based on the QCB-based DWT engine, three code blocks con be completely generated after every fixed time slice recursively. Thus, the DWT and EBCOT processors can process simultaneously and the high computational EBCOT then has the higher parallelism of the JPEG2000 encoding system. By changing the output timing of the DWT process and parallelizing with EBCOT, the internal tile memory size can be reduced by a factor of 4. The memory access cycles between the internal tile memory and the code block memory also decrease with the smooth encoding flow(1).
Analysis and architecture design for high performance JPEG2000 coprocessor
(2004)
Wu, BF
;
Lin, CF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
The high throughput and less internal memory requirement are two key issues of integrating the JPEG2000 encoding system. For this bottleneck of integration, a QCB (quad code block)-based DWT method is proposed to achieve more efficient parallelism in the JPEG2000 coprocessor. Based on the QCB-based DWT engine, the code blocks can be completely generated after every fixed time slice, recursively. Thus, the overall JPEG2000 encoding system arrives higher parallelism between DWT and EBCOT processors and thus preserves the high throughput, as DWT does. By changing the output timing of the DWT process and parallelizing with EBCOT, the internal tile memory size can be reduced by a factor of 4. Moreover, with smoother encoding flow of overall encoding system, the memory access cycles between the internal tile memory and code block memory also decreases.
Beryllium-implanted P-type GaN with high carrier concentration
(2001-05-01)
Yu, CC
;
Chu, CF
;
Tsai, JY
;
Lin, CF
;
Lan, WH
We report the results of beryllium implantation in Mg-doped GaN to obtain a high hole concentration and lower activation energy for Mg. The metal organic chemical vapor deposition (MOCVD)-grown Mg-doped GaN samples were implanted with Be ions at two different energies of 50 keV and 150 keV and a dose of 10(14) cm(-2). The implemented samples were subsequently annealed at 1100 degreesC for 60 s. The annealed samples showed an increase of hole concentration by three orders of magnitude from the non-implanted value of 5.5 x 10(16) to 8.1 x 10(19) cm(-3) as determined by Hall measurement. The activation energy of Mg dopants for the implanted annealed samples estimated from the temperature dependence of the photoluminescence data is about 170 meV, which is nearly 30% lower than that for the as-grown samples.
Blind adaptive energy estimation for decorrelating decision-feedback CDMA multiuser detection using learning-type stochastic approximations
(1999-03-01)
Chang, PR
;
Lee, CC
;
Lin, CF
;
電信工程研究所
;
Institute of Communications Engineering
This paper investigates the application of linear reinforcement learning stochastic approximation to the blind adaptive energy estimation for a decorrelating decision-feedback (DDF) multiuser detector over synchronous code-division multiple-access (CDMA) radio channels in the presence of multiple-access interference (MAI) and additive Gaussian noise. The decision feedback incorporated into the structure of a linear decorrelating detector is able to significantly improve the weaker users' performance by canceling the MAI from the stronger users. However, the DDF receiver requires the knowledge of the received energies, In this paper, a new novel blind estimation mechanism is proposed to estimate all the users' energies using a stochastic approximation algorithm without training data. In order to increase the convergence speed of the energy estimation, a linear reinforcement learning technique is conducted to accelerate the stochastic approximation algorithms. Results show that our blind adaptation mechanism is able to accurately estimate an the users' energies even if the users of the DDF detector are not ranked properly. After performing the blind energy estimation and then reordering the users in a nonincreasing order, numerical simulations show that the DDP detector for the weakest user performs closely to the maximum likelihood detector, whose complexity grows exponentially with the number of users.
Characteristics of optical properties of the interrupt growth method on InGaN/GaN MQW structures
(2000)
Lin, CF
;
Shu, CK
;
Lee, WH
;
Wen, TC
;
Chu, CF
The growth method and material properties of InGaN/GaN multiple quantum well (MQW) structures were studied in this experiment. The InGaN/GaN-MQW structure were deposited above the 2 mum-thick high quality GaN layers (with background carrier concentration was about 3.2 x 10(16) cm(-3)) on Al2O3 substrates. The interrupt growth time was 10 sec between GaN barrier layer and InGaN well growth at 690 degreesC. The stronger intensity and sharper photoluminescence (PL) spectra was observed from the interrupt growth MQW structures. The mean deviation of the PL peak position was lower than 1% by using the interrupt growth MQW structures at difference growth temperature. The PL spectra was become broaden and weaker by reducing the interrupt time during InGaN/GaN growth. The thickness of one pair InGaN/GaN was also identified by DCX-ray measurement and the growth rate InGaN layer was also calculated in this experiment.
Characteristics of stable emission GaN-based resonant-cavity light-emitting diodes
(2004-01-19)
Lin, CF
;
Yao, HH
;
Lu, JW
;
Hsieh, YL
;
Kuo, HC
The resonant-cavity light-emitting diode (RC-LED) structure was grown by MOCVD. The structure of RC-LED consisted of a 3lambda InGaN/GaN MQW LED cavity between the top TiO2/SiO2 DBR (81.7%, reflectance) and the bottom AlN/GaN DBR (90.4%) stack. A stable 410 nm emission peak and a low thermally induced red-shift effect (0.12 nm/kA/cm(2)) were measured by varying the injection current density. The light output power of the full RC-LED device was three times higher than the RC-LED without top TiO2/SiO2 DBR layers under 600 A/cm(2) inject current density. The narrow line width of 7.4 nm, emission peak localization at 410 nm, and three times higher output power were caused by the resonance effect in this vertical cavity structure. (C) 2003 Published by Elsevier B.V.
Characterization of GaN epitaxial layers on SiC substrates with AlxGa1-xN buffer layers
(1997-12-18)
Lin, CF
;
Cheng, HC
;
Feng, MS
;
Chi, GC
;
材料科學與工程學系
High quality GaN epitaxial layers were obtained with AlxGa1-xN buffer layers on 6H-SiC substrates. The low-pressure metalorganic chemical vapor deposition (LP-MOCVD) method was used. The 500 Angstrom thick buffer layers of AlxGa1-xN (0 less than or equal to x less than or equal to 1) were deposited on SiC substrates at 1025 degrees C. The FWHM of GaN (0004) X-ray curves are 2-3 arcmin, which vary with the Al content in AlxGa1-xN buffer layers. An optimum Al content is found to be 0.18. The best GaN epitaxial film has the mobility and carrier concentration about 564 cm(2) V-1 s(-1) and 1.6 x 10(17) cm(-3) at 300 K. The splitting diffraction angle between GaN and AlxGa1-xN were also analyzed from X-ray diffraction curves. (C) 1997 Elsevier Science S.A.
Chemical-mechanical polishing and material characteristics of plasma-enhanced chemically vapor deposited fluorinated oxide thin films
(1997-03-01)
Tseng, WT
;
Hsieh, YT
;
Lin, CF
;
Tsai, MS
;
Feng, MS
The chemical-mechanical polishing (CMP) process has been proven to be the most promising method for accomplishing global planarization. In this paper, results of chemical-mechanical polishing of fluorinated silicon dioxide (SiOF) thin films are presented. Nanohardness, elastic modulus, and bonding structure of fluorinated silicon dioxides are characterized in order to evaluate their correlations with CMP performance. The results show that under fixed chemistry and mechanical parameters, the CMP removal rate increases significantly with increasing fluorine content in the oxides due to the lower hardness and elastic modulus in the films. Higher CMP removal rate is observed for fluorinated oxides polished with slurry of pH 10 relative to pH 9. Compared with undoped oxides, SiOF films are more sentisitive to chemical and moisture attacks as reflected by the post-CMP increase in refractive index.
CMP of fluorinated silicon dioxide: Is it necessary and feasible?
(1997-02-01)
Tseng, WT
;
Hsieh, YT
;
Lin, CF
;
交大名義發表
;
National Chiao Tung University
Use of low-permittivity (low-k) fluorinated silicon dioxides (F-SiO2) for inter-metal dielectric (IMD) applications makes it necessary to evaluate the compatibility of chemical-mechanical polishing (CMP) in device fabrication. Basic material characteristics (such as adhesion, gap fill, step coverage, stress, hardness, chemical stability, moisture permeation, and wet etching rate) should be investigated thoroughly to assess CMP performance. With increasing fluorine content in the oxides, the low-k dielectric exhibits increased chemical reactivity and water absorption. These properties enhance the CMP removal rate, but degrade post-CMP material reliability and electrical performance. This article examines the CMP performance of F-SiO2 used to provide a low dielectric constant in IMD applications.
Comparison of p-side down and p-side up GaN light-emitting diodes fabricated by laser lift-off
(2003-02-15)
Chu, CF
;
Yu, CC
;
Cheng, HC
;
Lin, CF
;
Wang, SC
The performance characteristics of laser lift-off (LLO) freestanding InGaN/GaN multiple quantum well (MQW) light-emitting diodes (LEDs) mounted on a Cu substrate with p-side up, and p-side down configurations were determined and compared. The InGaN/GaN MQW LED structures, which were originally fabricated on a sapphire substrate, were transferred to a Cu substrate by the LLO process into two different configurations, namely p-side up and p-side down with the same Ni/Pd/Au p-contact metallizations. Both p-side down and p-side up LLO-LEDs showed a higher current operation capability up to 400 mA than the original LEDs on the sapphire substrate. The p-side down LLO-LEDs showed a nearly eightfold increase in the light output power compared to the p-side up LLO-LEDs. The p-side down LLO-LEDs also showed a more stable wavelength emission spectrum than the p-side up ones.
Correlation between dewatering index and dewatering performance of three mechanical dewatering devices
(2003-05-01)
Pan, JRS
;
Huang, CP
;
Cherng, MY
;
Li, KC
;
Lin, CF
Efficient sludge conditioning can improve sludge dewatering characteristics and promote the separation of flocs from the liquid phase to achieve a high solid content. To optimize this process, a reliable dewatering index must be established. Although numerous researches investigated sludge conditioning, few are devoted to the correlation between the conditioning index and the dewatering efficiency. In this study, sludges were conditioned with both synthetic and natural polymers, and their dewatering characteristics after three different mechanical dewatering devices, the filter press, belt press, and the centrifuge, were compared. The result shows that the traditionally used dewatering index may not reflect sludge conditioning properly. Correlation between sludge conditioning indices and dewaterability depends upon both the type of sludge and the means of sludge conditioning and dewatering. It is therefore concluded that dewatering index must be properly chosen for optimum sludge conditioning. (C) 2002 Elsevier Science Ltd. All rights reserved.
Design of multicode CDMA systems for 3-D stereoscopic video over wireless ATM networks
(2000-03-01)
Chang, PR
;
Lin, CF
;
Wu, MJ
;
電信工程研究所
;
Institute of Communications Engineering
This paper investigates the application of multicode spread-spectrum code-division multiple-access (SS-CDMA) techniques to three-dimensional (3-D) stereoscopic video transmission over wireless asynchronous transfer mode (ATM) networks. Three-dimensional visual communications, made through the use of stereoscopic images, are able to achieve total display realism. Such services allow users to share the virtual reality (VR) world without any geographical restrictions. In order to create a 3-D system with two images (left and right) that should be transmitted over a bandlimited mobile channel simultaneously, a cost-effective Motion Picture Experts Group (MPEG)-based wavelet multiresolution coding with a joint motion and disparity compensation is developed to reduce a large amount of information contained in the images to meet the low-transmission rate limitation of mobile channels. However, the rapidly variable bit rate (VBR) characteristics of the MPEG-based 3-D videos seems a weakness to the transmission of such videos via a constant bit-rate (CBR) mobile channel. The ATM technique is especially well suited for VER MPEG-based 3-D video because of its ability to allocate bandwidth on demand to these services. However, since the mobile radio has a Limited channel capacity, the overall capacity of the traditional ATM-based SS-CDMA system may not be sufficient to accommodate the MPEG-based 3-D video services requested by the multiple mobile users simultaneously. To tackle this difficulty, a multicode CDMA technique is proposed to provide VER MPEG-based 3-D video services by varying the number of spreading codes assigned to the 3-D video to meet its dynamic throughput requirement. Powerful forward error correction (FEC) codes are necessary to protect the video data so that it can be successfully transmitted at an acceptable signal power level. Two separate FEC code schemes are applied to the header and payload of an ATM cell containing 3-D video data, respectively. The ATM cell header is protected by a relatively powerful FEC code to ensure correct delivery and low cell loss rate (CLR). On the other hand, the ATM cell payload is encoded for varying degrees of error protection according to the priority of the payload data in 3-D videos. An adaptive FEC code combining scheme is proposed to provide the good protection for payload data with the maximization of its code rate to minimize the extra bandwidth for FEC overhead.
Design of spread spectrum multicode CDMA transport architecture for multimedia services
(2000-01-01)
Chang, PR
;
Lin, CF
;
電信工程研究所
;
Institute of Communications Engineering
In this paper, we investigate a new application of the well-known spread spectrum code division multiple access (SS-CDMA) techniques to multimedia services related to the development of the next-generation wireless mobile networks interconnecting with a wireline ATM-based broadband network. Such services allow users to share novel multimedia applications without any geographical restrictions. However, since the mobile radio channel has a fixed limited bandwidth, the traditional SS-CDMA system may not be sufficient to accommodate the variable bit rate (VBR) multimedia services requested by multiple mobile users simultaneously. Moreover, the traffic load at the base station can change dynamically due to the time-varying throughput requirement of these requested multimedia services; To tackle this difficulty, a multicode CDMA (MC-CDMA) technique is proposed to provide multirate multimedia services by varying the number of spreading codes assigned to each user in order to meet its throughput requirement. In MC-CDMA, a spreading code can be used to transmit information at a basic bit rate. Users (video or data) who need higher transmission rates can use multiple codes in parallel. Meanwhile, the maximum available number of codes in the MC-CDMA system is still limited. Hence, a cost-effective dynamic code allocation scheme has then been proposed to dynamically assign appropriate number of codes to each user for achieving the maximum resource utilization for multiuser multimedia services via the mobile radio channel, Finally, a number of real multimedia titles generated from the well-known MacroMind Director are conducted to evaluate the performance of the proposed wireless multimedia CDMA system.