• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.1121-1141
• /
• 2020

We propose a robust visual object tracking algorithm fusing a convolutional neural network tracker trained offline from a large number of video repositories and a color histogram based tracker to track objects for mixing immersive audio. Our algorithm addresses the problem of occlusion and large movements of the CNN based GOTURN generic object tracker. The key idea is the offline training of a binary classifier with the color histogram similarity values estimated via both trackers used in this method to opt appropriate tracker for target tracking and update both trackers with the predicted bounding box position of the target to continue tracking. Furthermore, a histogram similarity constraint is applied before updating the trackers to maximize the tracking accuracy. Finally, we compute the depth(z) of the target object by one of the prominent unsupervised monocular depth estimation algorithms to ensure the necessary 3D position of the tracked object to mix the immersive audio into that object. Our proposed algorithm demonstrates about 2% improved accuracy over the outperforming GOTURN algorithm in the existing VOT2014 tracking benchmark. Additionally, our tracker also works well to track multiple objects utilizing the concept of single object tracker but no demonstrations on any MOT benchmark.

• Journal of Korea Multimedia Society
• /
• v.20 no.2
• /
• pp.188-195
• /
• 2017

The paper presents a histogram equalization method using the edge information of an image to be processed. The basic idea of this method is to carry out histogram equalization with edge information, which is important and essential for object conformation. In the proposed method, the edge information is used to generate histogram for the equalization process. It is found to be effective to suppress the histogram spikes that cause quantum jumps in mapping function for the equalization process. The proposed method is tested for randomly selected 30 images and compared to conventional approaches with a quantitative measure to check it preserves the structural similarity. Experimental results show that the proposed method has better performance and no artifacts caused by histogram spikes.

• Korean Journal of Remote Sensing
• /
• v.31 no.1
• /
• pp.1-9
• /
• 2015

Mobility and cloud platform have become the dominant paradigm to develop web services dealing with huge and diverse digital contents for scientific solution or engineering application. These two trends are technically combined into mobile cloud computing environment taking beneficial points from each. The intention of this study is to design and implement a mobile cloud application for remotely sensed image fusion for the further practical geo-based mobile services. In this implementation, the system architecture consists of two parts: mobile web client and cloud application server. Mobile web client is for user interface regarding image fusion application processing and image visualization and for mobile web service of data listing and browsing. Cloud application server works on OpenStack, open source cloud platform. In this part, three server instances are generated as web server instance, tiling server instance, and fusion server instance. With metadata browsing of the processing data, image fusion by Bayesian approach is performed using functions within Orfeo Toolbox (OTB), open source remote sensing library. In addition, similarity of fused images with respect to input image set is estimated by histogram distance metrics. This result can be used as the reference criterion for user parameter choice on Bayesian image fusion. It is thought that the implementation strategy for mobile cloud application based on full open sources provides good points for a mobile service supporting specific remote sensing functions, besides image fusion schemes, by user demands to expand remote sensing application fields.

• Journal of Korea Multimedia Society
• /
• v.18 no.2
• /
• pp.138-148
• /
• 2015

Object detection methods based on background learning are widely used in video surveillance. However, when a car runs with headlights on, these methods are likely to detect the car region and the area illuminated by the headlights as one connected change region. This paper describes a method of separating the car region from the area illuminated by the headlights. First, we detect change regions with a background learning method, and extract blobs, connected components in the detected change region. If a blob is larger than the maximum object size, we extract candidate object regions from the blob by clustering the intensity histogram of the frame difference between the mean of background images and an input image. Finally, we compute the similarity between the mean of background images and the input image within each candidate region and select a candidate region with weak similarity as an object region.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.6
• /
• pp.2807-2814
• /
• 2011

This paper proposes an identification algorithm that detects detrimental digital video contents based on the color space features. In this paper, discrimination algorithm based on a 2-Dimensional Projection Maps is suggested to find targeted video images. First, 2-Dimensional Projection Maps which is extracting the color characteristics of the video images is applied to extract effectively detrimental candidate frames from the videos, and next estimates similarity between the extracted frames and normative images using the suggested algorithm. Then the detrimental candidate frames are selected from the result of similarity evaluation test which uses critical value. In our experimental test, it is suggested that the results of the comparison between the Color Histogram and the 2-Dimensional Projection Maps technique to detect detrimental candidate frames. Through the various experimental data to test the suggested method and the similarity algorithm, detecting method based on the 2-Dimensional Projection Maps show more superior performance than using the Color Histogram technique in calculation speed and identification abilities searching target video images.

• Journal of Internet Computing and Services
• /
• v.9 no.5
• /
• pp.117-130
• /
• 2008

This paper proposes a stable and robust method to distinct the color similarity for clothes using the LBG algorithm under various light sources, Since the conventional methods, such as the histogram intersection and the accumulated histogram, are profoundly sensitive to the changing of light environments, the distinction of color similarity for the same cloth can be different due to the complicated light sources. To reduce the effects of the light sources, the properties of hue and saturation which consistently sustain the characteristic of the color under the various changes of light sources are analyzed to define the characteristic of the color distribution. In a two-dimensional space determined by the properties of hue and saturation, the LBG algorithm, a non-parametric clustering approach, is applied to examine the color distribution of images for each clothes. The color similarity of images is defined by the average of Euclidean distance between the mapping clusters which are calculated from the result of clustering of both images. To prove the stability of the proposed method, the results of the color similarity between our method and the traditional histogram analysis based methods are compared using a dozen of cloth examples that obtained under different light environments. Our method successively provides the classification between the same cloth image pair and the different cloth image pair and this classification of color similarity for clothe images obtains the 91.6% of success rate.

• Journal of information and communication convergence engineering
• /
• v.11 no.1
• /
• pp.62-68
• /
• 2013

This paper proposes a method for identification of harmful video images based on the degree of harmfulness in the video content. To extract harmful candidate frames from the video effectively, we used a video color extraction method applying a projection map. The procedure for identifying the harmful video has five steps, first, extract the I-frames from the video and map them onto projection map. Next, calculate the similarity and select the potentially harmful, then identify the harmful images by comparing the similarity measurement value. The method estimates similarity between the extracted frames and normative images using the critical value of the projection map. Based on our experimental test, we propose how the harmful candidate frames are extracted and compared with normative images. The various experimental data proved that the image identification method based on the 2-dimensional projection map is superior to using the color histogram technique in harmful image detection performance.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.1
• /
• pp.1-7
• /
• 2002

In this paper, when users choose a query image, we implemented a content-based image retrieval system that users can simply choose and extract a object region of query wanted with not only a whole image but various objects in it. Histogram is obtained by improved HSV transformations from query image and then candidate images are retrieved rapidly by a 1st similarity measure with histogram intersection using representative colors of query image. And finally retrieved images are extracted since 2nd similarity measure with banded autocorrelogram is performed so that recall and precision are improved by combining two retrieval methods that can make up for respective weak points. Moreover images in the database are indexed automatically within feature library that makes possible to retrieve images rapidly.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.4
• /
• pp.548-553
• /
• 2010

This paper presents an effective shape image recognition method using the correlation based on 4-dimensional histogram. The histogram-based correlation is accurately applied to express the similarity by comparing the positions of a corresponding dimension between the images, which is calculated by considering 4 directions of the shape image. The correlation measure by using the normalized cross-correlation is also applied to obtain the robust recognition to the geometrical variations such as shape, position, size, and rotation. The proposed method has been applied to the problem for recognizing the 8 shape images of 64*64 pixels and the 30 shape images of 256*256 pixels. The experimental results show that the proposed method has a superior recognition performance that appears the image characters well.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.705-708
• /
• 2010

An eigen color co-occurrence approach is proposed that exploits the correlation between color channels to identify the degree of image similarity. This method is based on traditional co-occurrence matrix method and histogram equalization. On the purpose of feature extraction, eigen color co-occurrence matrices are computed for extracting the statistical relationships embedded in color images by applying Principal Component Analysis (PCA) on a set of color co-occurrence matrices, which are computed on the histogram equalized images. That eigen space is created with a set of orthogonal axes to gain the essential structures of color co-occurrence matrices, which is used to identify the degree of similarity to classify an input image to be tested for various purposes. In this paper RGB, Gaussian color space are compared with grayscale image in terms of PCA eigen features embedded in histogram equalized co-occurrence features. The experimental results are presented.