Hui-Yu Huang

Color image segmentation using a self-organizing map algorithm

A color image segmentation methodology based on a self-organizing map (SOM) is proposed. The method developed takes into account the color similarity and spatial relationship of ob- jects within an image. According to the features of color similarity, an image is first segmented into coarse cluster regions. The resulting regions are then treated by computing the spatial distance between any two cluster regions, and the SOM with a labeling process is applied. In this paper, the selection of the parameters for the SOM algorithm was also investigated experimentally. The experimental results show that the proposed system is feasible, and that the seg- mented object regions are similar to those perceived by human vi- sion.

An embedded watermark technique in video for copyright protection

As the Internet and storage equipment become more and more populous, the digital data can easy to distribution. The pirate can easy and limitless to copy digital multimedia. Therefore people are more and more attentive the copyright protection for digital multimedia. Digital watermarking is an important emerging technique for copyright protection and authentication. This paper presents a novel video-watermarking algorithm based on block matching algorithm to achieve copyright protection of owner. The watermark is mainly embedded in the uncompressed domain and is detected without the use of the original video information. Our proposed method is easy to perform and can provide an adjustable threshold to obtain the imperceptibility and robustness performance. The obtained results show the robustness of this approach

A Film Classifier Based on Low-level Visual Features

In this paper, we propose an approach to categorize the film classes by using low-level features and visual features. The goal of this approach is to classify the films into genres. Our current domain of study is using the movie preview. A film preview often emphasizes the theme of a film and hence provides suitable information for classification process. In our approach, we classify films into three broad categories: action, dramas, and thriller films. Four computable video features (average shot length, color variance, motion content and lighting key) and visual effects are combined in our approach to provide the advantage information to demonstrate the movie category. Our approach can also be extended for other potential applications, including the browsing and retrieval of videos on the Internet, video-on-demand, and video libraries.

An adaptive video watermarking technique based on DCT domain

In this paper, we propose an effective video embedded watermarking technique based on DCT domain with high transparency and slight distortion. The watermark is mainly embedded into the uncompressed domain by adjusting the correlation between DCT coefficients of the selected blocks, and it can be detected without the original video data. The system contains the preprocessing, watermark embedding and watermark extraction. In the preprocessing, in order to improve the computation complexity and reduce the computation time, a pseudo 3D DCT by two times of the DCT transformation will firstly be obtained. In the embedding process, we embed the watermark into the successive raw frames transformed into the pseudo codes before compression, afterward a secret embedding key will be created. This secret embedding key will further use to the extraction processing. The experimental results show that the proposed scheme is extremely robust to against various attacks.

Movie Classification Using Visual Effect Features

In this paper, we propose an approach to category the film kinds using low-level features and visual effect features. This approach aims to category the films into genres. Our current domain of study is the movie preview. A film preview often emphasizes the theme of a film and hence provides suitable information for classification process. In our approach, we classify films into three broad categories: Action, Dramas, and Thriller films. Four computable video features (average shot length, color variance, motion content and lighting key) and visual effects are combined in our approach to provide the advantage information to demonstrate the movie category. Our approach can also be extended for other potential applications, including browsing and retrieval of videos on the internet, video-on-demand, and video libraries. Experimental results show the visual effect feature play an important role for classifying the film kinds.

A video watermarking algorithm based on pseudo 3D DCT

In this paper, we propose an adaptive video watermarking algorithm based on a pseudo 3D DCT to insert the high transparency and slight distortion messages to resist the attacks. The watermark is mainly inserted into the uncompressed domain by adjusting the correlation between DCT coefficients of the selected blocks, and the watermark extraction is blind, i.e., no original unwatermarked video is needed for watermark extraction. The system consists of pseudo 3D DCT technique, watermark embedding, and extraction. A pseudo 3D DCT technique will utilize to calculate the embedding factor and the advantageous messages. In the embedding process, using the quantization index modulation (QIM), we embed the watermark into the quantization regions from the successive raw frames in the uncompressed domain and record the relative information to create a secret embedding key. This secret embedding key will further apply to watermark extraction. Experimental results show that the proposed method can obtain the good performance in transparency and robustness against various attacks such as filtering, compression, and addition of noise.

Primary‐view perception on a gray image: Region segmentation and association

Abstract The segmentation of scenes into perceptually meaningful partitions possessing a useful relationship among them is of great importance in image understanding. In this paper, an effective approach performing image segmentation and association among the segmented regions for a gray image is presented. An unsupervised segmentation process using a self‐organization map (SOM) and spatial‐distance computation is presented for the segmentation of clusters in our understanding system. An association process is developed for the construction of spatial relationships among the attended regions, which are obtained from the segmented clusters using fuzzy relations. The algorithms for each process are presented and exemplified with a series of illustrations. This approach is applied to simulating the primary‐view perception on a natural gray image. Experiments show that the proposed approach is feasible. A possible linguistic presentation, based on the attended regions and their relationships, is given for every test image.

Film classification based on low-level visual effect features

We present a framework to classify the film categories based on low-level features and visual effect features. This ap- proach can serve as a prefilter that represents the movies that you want to watch from the Internet or movie-on-demand service. Our current domain of study is the movie preview. In our approach, we categorize films into three broad kinds: action, drama, and thriller films. Four low-level video features (average shot length, color vari- ance, lighting key, and motion content) and visual effects are com- bined in our system to provide helpful information to demonstrate the film category. The results indicate that visual effect features are effective and useful information and play an important role for clas- sifying the film categories.

A movie classifier based on visual features

In this paper, we propose an approach to classify the film categories by using low-level features and visual features. The goal of this approach is to classify the films into genres. Our current domain of study is the movie preview. A film preview often emphasizes the theme of a film and hence provides suitable information for classification process. In our approach, we classify films into three broad categories: Action, Dramas, and Thriller films. Four computable video features (average shot length, color variance, motion content and lighting key) and visual effects are combined in our approach to provide the advantage information to demonstrate the movie category. Our approach can also be extended for other potential applications, including browsing, retrieval of videos on the internet, video-on-demand, and video libraries.

Integrating color, texture, and spatial features for image interpretation

In this paper, we present an approach to achieve the region-based image semantic interpretation and recall process in color image from image database. This system includes feature extraction of region, indexing process, linguistic inference rules construction, as well as a semantic description of region image. Based on these features, each of human labeled regions in an image can be described by a corresponding linguistic meaning. The main procedure consists of two parts: procedure 1 (forward) and 2 (recall) processes. The forward process primarily presents the linguistic meaning description of a region image based on feature definitions, inference rules, and indexing process. In recall process, it mainly reconstructs the region image which performs the rough mental image of human memory retrieval according to the semantic meaning by means of a specified or the pre-staged result. Experiments confirm that our approach is reasonable and feasible.