• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.8
• /
• pp.3790-3803
• /
• 2018

Content-based image retrieval is an approach used to query images based on their semantics. Semantic based retrieval has its application in all fields including medicine, space, computing etc. Semantically generated binary hash codes can improve content-based image retrieval. These semantic labels / binary hash codes can be generated from unlabeled data using convolutional autoencoders. Proposed approach uses semi-supervised deep hashing with semantic learning and binary code generation by minimizing the objective function. Convolutional autoencoders are basis to extract semantic features due to its property of image generation from low level semantic representations. These representations of images are more effective than simple feature extraction and can preserve better semantic information. Proposed activation and loss functions helped to minimize classification error and produce better hash codes. Most widely used datasets have been used for verification of this approach that outperforms the existing methods.

• Journal of KIISE:Software and Applications
• /
• v.31 no.3
• /
• pp.332-342
• /
• 2004

With the explosive growth of information in our real life, information filtering is quickly becoming a popular technique for reducing information overload. Information filtering technique is divided into two categories: content-based filtering and collaborative filtering (or social filtering). Content-based filtering selects the information based on contents; while collaborative filtering combines the opinions of other persons to make a prediction for the target user. In this paper, we describe a new filtering approach that seamlessly combines content-based filtering and collaborative filtering to take advantages from both of them, where a technique using user profiles efficiently on the collaborative filtering framework is introduced to predict a user's preference. The proposed approach is experimentally evaluated and compared to conventional filtering. Our experiments showed that the proposed approach not only achieved significant improvement in prediction quality, but also dealt with new users well.

• Speech Sciences
• /
• v.12 no.1
• /
• pp.135-142
• /
• 2005

An important aspect of video indexing is the ability to segment video into meaningful segments, i.e., content-based video segmentation. Since the audio signal in the sound track is synchronized with image sequences in the video program, a speech signal in the sound track can be used to segment video into meaningful segments. In this paper, we propose a new approach to content-based video segmentation. This approach uses closed caption to construct a recognition network for speech recognition. Accurate time information for video segmentation is then obtained from the speech recognition process. For the video segmentation experiment for TV news programs, we made 56 video summaries successfully from 57 TV news stories. It demonstrates that the proposed scheme is very promising for content-based video segmentation.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.2
• /
• pp.16-20
• /
• 2001

An important aspect of video program indexing and retrieval is the ability to segment video program into meaningful segments, in other words, the ability of content-based video program segmentation. In this paper, a new approach using speech recognition technology has been proposed for content-based video program segmentation. This approach uses speech recognition technique to synchronize closed caption with speech signal. Experimental results demonstrate that the proposed scheme is very promising for content-based video program segmentation.

• The Journal of Society for e-Business Studies
• /
• v.26 no.3
• /
• pp.119-132
• /
• 2021

Identification of associated knowledge based on content relevance is a fundamental functionality in managing service and security of core knowledge. This study compares the performance of methods to identify associated knowledge based on content relevance, i.e., the associated document network composition performance of keyword-based and word-embedding approach, to examine which method exhibits superior performance in terms of quantitative and qualitative perspectives. As a result, the keyword-based approach showed superior performance in core document identification and semantic information representation, while the word embedding approach showed superior performance in F1-Score and Accuracy, association intensity representation, and large-volume document processing. This study can be utilized for more realistic associated knowledge service management, reflecting the needs of companies and users.

• International Journal of Contents
• /
• v.4 no.2
• /
• pp.24-28
• /
• 2008

In this paper, we propose a new approach for content-based video retrieval using non-parametric based motion classification in the shot-based video indexing structure. Our system proposed in this paper has supported the real-time video retrieval using spatio-temporal feature comparison by measuring the similarity between visual features and between motion features, respectively, after extracting representative frame and non-parametric motion information from shot-based video clips segmented by scene change detection method. The extraction of non-parametric based motion features, after the normalized motion vectors are created from an MPEG-compressed stream, is effectively fulfilled by discretizing each normalized motion vector into various angle bins, and by considering the mean, variance, and direction of motion vectors in these bins. To obtain visual feature in representative frame, we use the edge-based spatial descriptor. Experimental results show that our approach is superior to conventional methods with regard to the performance for video indexing and retrieval.

• ETRI Journal
• /
• v.36 no.5
• /
• pp.714-720
• /
• 2014

In this paper, we propose the use of data-driven probabilistic utterance-level decision logic to improve Weighted Finite State Transducer (WFST)-based endpoint detection. In general, endpoint detection is dealt with using two cascaded decision processes. The first process is frame-level speech/non-speech classification based on statistical hypothesis testing, and the second process is a heuristic-knowledge-based utterance-level speech boundary decision. To handle these two processes within a unified framework, we propose a WFST-based approach. However, a WFST-based approach has the same limitations as conventional approaches in that the utterance-level decision is based on heuristic knowledge and the decision parameters are tuned sequentially. Therefore, to obtain decision knowledge from a speech corpus and optimize the parameters at the same time, we propose the use of data-driven probabilistic utterance-level decision logic. The proposed method reduces the average detection failure rate by about 14% for various noisy-speech corpora collected for an endpoint detection evaluation.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2006.05b
• /
• pp.197-200
• /
• 2006

The ingress of chloride ions plays a crucial role for service life design of reinforced concrete structures. In view of durability design of concrete structures under marine environment, one of the most essential parameters is the surface chloride content of concrete. However, on the basis of the results of in-situ investigation, this value has been determining in the numerous studies on the durability design of concrete structures. Hence, it is necessary to confirm the range of the surface chloride content in order to establish a unified durability design system of concrete. This study suggests a rational and practical way to calculate the maximum surface chloride content of submerged concrete under marine environment. This approach starts with the calculation of the amount of chloride ingredients in normal sea water. The capillary pore structure is modeled by numerical simulation model HYMOSTRUC and it is assumed to be completely saturated by the salt ingredients of sea water. In order to validate this approach, the total chloride content of the mortar and concrete slim disc specimen was measured after the immersion into the artificial sea water solution. Additionally, the theoretical, the experimental and in-situ investigation results of other researchers are compiled and analyzed. Based on this approach, it will follow to calculate the maximum surface chloride content of concrete at tidal zone, where the environment can be considered as a condition of dry-wetting cycles.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.12 no.1
• /
• pp.50-59
• /
• 1987

So far a variety of RAM-based approaches including the Filtering Method have been suggested to shorten the rule seletion time in production systems, but this paper presents a somewhat different approach based on the use of CAM. This paper suggests a proper use of CAM bits respect to their characteristics and describes data stsuctures for basic Artificial Intelligence symbolic list processing, and finally compares the simulation results from the CAM-based approach to those from RAM-based approaches.

• Journal of the HCI Society of Korea
• /
• v.2 no.1
• /
• pp.49-56
• /
• 2007

A nonintrusive projection-based AR approach using complementary pattern has been recently proposed and applied to virtual studio. However, the approach faces the tradeoff between the pattern imperceptibility and compensation accuracy. To alleviate the tradeoff, we propose a content adaptive pattern concealment approach. The projector input images (AR images) are divided into rectangular regions and spatial variation and color distribution are computed in the regions. Based on the spatial variation and color distribution, we embed locally different strength of pattern images into different color channels. It is demonstrated that the proposed approach has two opposite advantages by comparing it with the previous (non-adaptive) approach through a variety of experiments and subjective evaluation. Our content adaptive approach can obtain the same performance using weaker pattern than the previous approach and thus significantly improve the imperceptibility of the pattern. On the contrary, our content adaptive approach can make strong pattern less perceptible and thus produce better compensation results.