Browsing by Author "Tsai, Yi-Hsuan"

Now showing 1 - 5 of 5

An RFID application in the food supply chain: A case study of convenience stores in Taiwan
(2011-09-01)
Hong, I-Hsuan
;
Dang, Jr-Fong
;
Tsai, Yi-Hsuan
;
Liu, Chen-Shen
;
Lee, Wang-Tsang
Food hazards can appear at any stage of global food supply chains, making it essential to define critical control points to capture the data about ingredients, manufacture and dates-certain (sell-by, use-by), etc., and provide it in a transparent manner to supply chain participants and consumers. The government of Taiwan has appointed a non-profit research organization to conduct a pilot project to launch a potential national-wide food traceability system to increase the intangible value of purchased food and to enhance food safety. This paper discusses a financially viable business model for a Radio Frequency Identification (REID) application to a food traceability system. We conduct a case study of REID implementation in the chain of convenience stores in Taiwan. The Taiwanese experiment may have implications for policy-makers, industry and public health officials elsewhere. (C) 2011 Elsevier Ltd. All rights reserved.
Bridging Stereo Matching and Optical Flow via Spatiotemporal Correspondence
(2019-01-01)
Lai, Hsueh-Ying
;
Tsai, Yi-Hsuan
;
Chiu, Wei-Chen
;
交大名義發表
;
National Chiao Tung University
Stereo matching and flow estimation are two essential tasks for scene understanding, spatially in 3D and temporally in motion. Existing approaches have been focused on the unsupervised setting due to the limited resource to obtain the large-scale ground truth data. To construct a self-learnable objective, co-related tasks are often linked together to form a joint framework. However, the prior work usually utilizes independent networks for each task, thus not allowing to learn shared feature representations across models. In this paper, we propose a single and principled network to jointly learn spatiotemporal correspondence for stereo matching and flow estimation, with a newly designed geometric connection as the unsupervised signal for temporally adjacent stereo pairs. We show that our method performs favorably against several state-of-the-art baselines for both unsupervised depth and flow estimation on the KITTI benchmark dataset.
Plug-and-Play: Improve Depth Prediction via Sparse Data Propagation
(2019-01-01)
Wang, Tsun-Hsuan
;
Wang, Fu-En
;
Lin, Juan-Ting
;
Tsai, Yi-Hsuan
;
Chiu, Wei-Chen
We propose a novel plug-and-play (PnP) module for improving depth prediction with taking arbitrary patterns of sparse depths as input. Given any pre-trained depth prediction model, our PnP module updates the intermediate feature map such that the model outputs new depths consistent with the given sparse depths. Our method requires no additional training and can be applied to practical applications such as leveraging both RGB and sparse LiDAR points to robustly estimate dense depth map. Our approach achieves consistent improvements on various state-of-the-art methods on indoor (i.e., NYU-v2) and outdoor (i.e., KITTI) datasets. Various types of LiDARs are also synthesized in our experiments to verify the general applicability of our PnP module in practice.
The inhibitory effect of 7, 7 ''-dimethoxyagastisflavone on the metastasis of melanoma cells via the suppression of F-actin polymerization
(2017-09-01)
Lin, Ching-Min
;
Lin, Yu-Ling
;
Ho, Shu-Yi
;
Chen, Pin-Rong
;
Tsai, Yi-Hsuan
7,7"-Dimethoxyagastisflavone (DMGF), a biflavonoid isolated from Taxus x media cv. Hicksii, induces apoptotic and autophagic cell death. However, whether DMGF suppresses tumor metastasis is unclear. The aim of this study was to investigate the anti-metastatic activities of DMGF on the metastatic processes of melanoma cells in vivo and in vitro. A transwell assay showed that DMGF could effectively attenuate the motility of B16F10 cells, and the results of real-time PCR revealed that DMGF also suppressed the expressions of matrix metalloproteinase-2 (MMP-2). Moreover, DMGF did not influence tube formation but inhibited the migration of endothelial cells. Furthermore, animal models were used to monitor the effects of DMGF on tumor metastasis, and all models showed that DMGF significantly suppressed the metastatic behaviors of B16F10 cells, including intravasation, colonization, and invasion of the lymphatic duct. In addition, DMGF could also reduce the densities of the blood vessels in the tumor area in vivo. Further investigation of the molecular mechanisms of anti-metastatic activity revealed that DMGF can down-regulate the levels of key modulators of the Cdc42/Rac1 pathway to interfere in F-actin polymerization and suppress the formation of lamellipodia by reducing the phosphorylation of CREB. These data suggested that DMGF presents anti-metastatic activities in B16F10 melanoma cells. Here, we demonstrated that DMGF can inhibit the metastasis of highly invasive melanoma cancer cells through the down-regulation of F-actin polymerization. Considering these findings, DMGF may be further developed to serve as a chemoprevention drug for patients with metastatic melanoma.
VOSTR: Video Object Segmentation via Transferable Representations
(1970-01-01)
Chen, Yi-Wen
;
Tsai, Yi-Hsuan
;
Lin, Yen-Yu
;
Yang, Ming-Hsuan
;
交大名義發表
In order to learn video object segmentation models, conventional methods require a large amount of pixel-wise ground truth annotations. However, collecting such supervised data is time-consuming and labor-intensive. In this paper, we exploit existing annotations in source images and transfer such visual information to segment videos with unseen object categories. Without using any annotations in the target video, we propose a method to jointly mine useful segments and learn feature representations that better adapt to the target frames. The entire process is decomposed into three tasks: (1) refining the responses with fully-connected CRFs, (2) solving a submodular function for selecting object-like segments, and (3) learning a CNN model with a transferable module for adapting seen categories in the source domain to the unseen target video. We present an iterative update scheme between three tasks to self-learn the final solution for object segmentation. Experimental results on numerous benchmark datasets demonstrate that the proposed method performs favorably against the state-of-the-art algorithms.