Browsing by Author "Wang, Tsai-Pei"

Now showing 1 - 2 of 2

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.
Multi-resolution volume rendering of large time-varying data using video-based compression
(2008)
Ko, Chia-Lin
;
Liao, Horng-Shyang
;
Wang, Tsai-Pei
;
Fu, Kuang-Wei
;
Lin, Ching-Yao
We present a new framework that combines the hierarchical multi-resolution representation with video-based compression to manage and render large scale time-varying data. In the preprocessing step, the proposed method first constructs a multi-resolution hierarchy using octree structure for each individual time step, and then applies a motion-compensation-based prediction to compress the octree nodes. During rendering stage, the data is decompressed on-the-fly and rendered using hardware texture mapping. The proposed approach eliminates the hierarchical decompression dependency commonly found in the conventional hierarchical wavelet representation methods, which leads to a more efficient reconstruction of data along the time axis. The system provides the user with a spatial region-of-interest (ROI) to adjust the spatial level-of-detail (LOD) selection, and a temporal ROI which is a sub-region only for frequent update during playback. With a suitable control of both ROIs, our system can reach an interactive playback frame rate. This allows the user to observe the dynamic nature of large time-varying data sets.