Browsing by Author "Lin, Fang-Ju"

Now showing 1 - 8 of 8

A micro-architecture simulator for multimedia stream processor
(2006)
Lin, Fang-Ju
;
Chiueh, Herming
;
電信工程研究所
;
Institute of Communications Engineering
Recent research has proposed using stream processors for media applications. Since a programmable Stream Processor could utilize various hardware micro-architectures for diverse media applications, decision on a suitable micro-architectures to achieve efficiencies and hardware cost is critical. In this paper, a micro-architecture simulator for stream processor is implemented. The simulator evaluate the performance of media application executed on various micro-architectures, and then to analyze the utility rate of hardware and consumption of memory. By comparing the performance of media application executed on diverse micro-architectures, the optimized hardware micro-architecture can be determined for specific application and suitable micro-architecture of Stream Processor can be implemented on later VLSI for different targeting systems.
Alpha matting using robust color sampling and fully connected conditional random fields
(2018-06-01)
Lin, Fang-Ju
;
Chuang, Jen-Hui
;
資訊工程學系
;
Department of Computer Science
Alpha matting refers to the problem of softly extracting the foreground from a given image. Previous matting approaches often focused on using na < ve color sampling methods to estimate foreground and background colors for unknown pixels. Existing sampling-based matting methods often collect samples only near the unknown pixels, which may yield poor results if the true foreground and background samples are not found. In this paper, we present novel approach to extract foreground elements from an image through color and opacity (i.e., alpha) estimations, which consider available samples in a search window of variable size for each unknown pixel. Our proposed sampling method is robust in that similar sampling results can be generated for input trimaps of different unknown regions. Further, after the initial estimation of the alpha matte, a fully connected conditional random field (CRF) is used to correct the predicted matte at the pixel level. Our experiments show that visually plausible alpha mattes can indeed be produced.
Directional Solution Coating by the Chinese Brush: A Facile Approach to Improving Molecular Alignment for High-Performance Polymer TFTs
(2017-09-13)
Lin, Fang-Ju
;
Guo, Cheng
;
Chuang, Wei-Tsung
;
Wang, Chien-Lung
;
Wang, Qianbin
Directional solution coating by the Chinese brush provides a facile approach to fabricate highly oriented polymer thin films by finely controlling the wetting and dewetting processes under directional stress. The biggest advantage of the Chinese brush over the normal western brush is the freshly emergent hairs used, whose unique tapered structure renders a dynamic balance of the liquid within the brush by multiple forces when interacting with the liquid. Consequently, the liquid is steadily held within the brush without any unexpected leakage, making the liquid transfer proceed in a well-controllable manner. It is demonstrated that the Chinese brush coating enables the crystallization of the polymer and the self-assembly of conjugated backbones to proceed in a quasi-steady state via a certain direction, which is attributed to the controllable receding of the three-phase contact line during the dewetting process by the multiple parallel freshly emergent hairs. The as-prepared polymer thin films exhibit over six times higher charge-carrier mobility compared to the spin-coated films, which therefore provides a general approach for high-performance organic thin-film transistors.
Image super-resolution by estimating the enhancement weight of self example and external missing patches
(2018-08-01)
Lin, Fang-Ju
;
Chuang, Jen-Hui
;
資訊工程學系
;
Department of Computer Science
Image super-resolution (SR) is the process of generating a high-resolution (HR) image using one or more low-resolution (LR) inputs. Many SR methods have been proposed, but generating the small-scale structure of an SR image remains a challenging task. We hence propose a single-image SR algorithm that combines the benefits of both internal and external SR methods. First, we estimate the enhancement weights of each LR-HR image patch pair. Next, we multiply each patch by the estimated enhancement weight to generate an initial SR patch. We then employ a method to recover the missing information from the high-resolution patches and create that missing information to generate a final SR image. We then employ iterative back-projection to further enhance visual quality. The method is compared qualitatively and quantitatively with several state-of-the-art methods, and the experimental results indicate that the proposed framework provides high contrast and better visual quality, particularly for non-smooth texture areas.
Metric learning for weather image classification
(2018-06-01)
Lin, Fang-Ju
;
Wang, Tsai-Pei
;
資訊工程學系
;
Department of Computer Science
Image classification is a core task in many applications of computer vision. Recognition of weather conditions based on large-volume image datasets is a challenging problem. However, there has been little research on weather-related recognition using color images, particularly with large datasets. In this study, we proposed a metric learning framework to investigate a two-class weather classification problem. We improve the classification accuracy using metric learning approaches. Extracting features from images is a challenging task and practical requirements such as domain knowledge and human intervention. In this paper, we define several categories of weather feature cures based on observations of outdoor images captured under different weather conditions. Experimental results show that a classifier based on metric learning framework is effective in weather classification and outperforms the previous approach when using the same dataset.
Novel conjugated polymers based on bis-dithieno [3,2-b;2 ',3 '-d] pyrrole vinylene donor and diketopyrrolopyrrole acceptor: side chain engineering in organic field effect transistors
(2018-01-07)
Lin, Fang-Ju
;
Lin, Song-Di
;
Chin, Chih-Hao
;
Chuang, Wei-Tsung
;
Hsu, Chain-Shu
Two new donor-acceptor conjugated polymers, i.e. PB(C12DTP) V-DTDPP-C-12 and PB(C12DTP) V-DTDPP-C12C8, were prepared by the Stille coupling of (E)-1,2-bis(4-dodecyl-6-(trimethylstannyl)-4H-dithieno[ 3,2-b: 2', 3'-d] pyrrol-2-yl) ethene (Sn-B(C12DTP)V) with 3,6-bis(5-bromothiophen-2-yl)-2,5-didodecylpyrrolo[ 3,4-c] pyrrole-1,4-dione (Br-DTDPP-C-12) and 3,6-bis(5-bromo-thiophen-2-yl)-2,5-bis(2octyldodecyl) pyrrolo[ 3,4-c] pyrrole-1,4-dione (Br-DTDPP-C8C12), respectively. The molecular weight and thermal properties of the obtained polymers were characterized by gel permeation chromatography (GPC), differential scanning calorimetry (DSC) and thermogravimetry (TG). Their optical and electrochemical properties were analyzed by UV-vis-NIR spectroscopy and cyclic voltammetry (CV). The morphological and molecular-packing features of the polymers were characterized by optical microscopy and atomic force microscopy (AFM), UV polarization spectroscopy and grazing incidence X-ray diffraction (GIXD) to determine the relationship between the structural morphology and the performance. PB (C12DTP) V-DTDPP-C12C8 with branched C12C8 side chains on DPP units shows much higher hole mobility (1.2 cm(2) V-1 s(-1)) than 0.69 cm(2) V-1 s(-1) of PB(C12DTP) V-DTDPP-C-12.
Syntheses and specific interactions of poly (hydroxyethyl methacrylate-b-vinyl pyrrolidone) diblock copolymers and comparisons with their corresponding miscible blend systems
(2006-09-20)
Huang, Chih-Feng
;
Kuo, Shiao-Wei
;
Lin, Fang-Ju
;
Wang, Chih-Feng
;
Hung, Chen-Jui
Combination of atom transfer radical and conventional free radical polymerizations has been successfully used to prepare poly(hydroxyethyl methacrylate-b-vinyl pyrrolidone) (PHEMA-b-PVP) copolymers with controlled molecular weight and low polydispersity (< 1.4). The thermal behavior and specific interaction of PHEMA-b-PVP diblock copolymers and their corresponding PHEMA/PVP blends were characterized. The result shows that glass transition temperatures of diblock copolymers analysed by differential scanning calorimetry (DSC) are higher than those of the blends. Infrared and solid-state NMR spectroscopic analyses show that hydrogen-bonding interaction of hydroxyl-carbonyl groups of diblock copolymers was also greater than that of the blends. Measurement of the proton spin-lattice relaxation time in the rotating frame, (T-1P(H)), reveals that all diblock copolymers and blends possess one composition-dependent T H, indicating that both diblock copolymers and blends are homogeneous, which is consistent with the DSC analysis. (c) 2006 Elsevier Ltd. All rights reserved.
Synthesis of a 4,9-Didodecyl Angular-Shaped Naphthodiselenophene Building Block To Achieve High-Mobility Transistors
(2016-07-26)
Tsai, Che-En
;
Yu, Ruo-Han
;
Lin, Fang-Ju
;
Lai, Yu-Ying
;
Hsu, Jhih-Yang
A new tetracyclic 4,9-dialkyl angular-shaped naphthodiselenophene (4,9-alpha-aNDS) was designed and synthesized. The naphthalene core in 4,9-alpha-aNDS is formed by the DBU-induced 6 pi-cyclization of an (E)-1,2-bis(3-(tetradec-1-yn-1-yl)selenophen-2-yl)ethene intermediate followed by the second PtCl2-catalyzed benzannulation. This synthetic protocol allows for incorporating two dodecyl groups regiospecifically at 4,9-positions of the resulting alpha-aNDS. An ordered supramolecular self-assembly formed via noncovalent selenium-selenium interactions with a short contact of 3.5 angstrom was observed in the single-crystal structure of 4,9-alpha-aNDS. The distannylated alpha-aNDS building block was copolymerized with Br-DTFBT and Br-DPP acceptors by Stille cross coupling to form two new donor-acceptor polymers P alpha NDSDTFBT and P alpha NDSDPP, respectively. The bottom-gate/top-contact organic field-effect devices using the PaNDSDTFBT and PaNDSDPP semiconductors accomplished superior hole mobility of 3.77 and 2.17 cm(2) V-1 s(-1), respectively, which are among the highest mobilities reported to date.