• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.306-309
• /
• 1996

This paper presents a method of landmark recognition in indoor environments using a neural-network for an autonomous mobile robot. In order to adapt to image deformation of a landmark resulted from variations of view-points and distances, a multi-labeled template matching(MLTM) method and a dynamic area search method(DASM) are proposed. The MLTM is. used for matching an image template with deformed real images and the DASM is proposed to detect correct feature points among incorrect feature points. Finally a feed-forward neural-network using back-propagation algorithm is adopted for recognizing the landmark.

• Interaction and multiscale mechanics
• /
• v.3 no.4
• /
• pp.373-387
• /
• 2010

This study is intended to assess low frequency sound radiated from a viaduct under normal traffic. The bridge comprises steel box girders and wide cantilever decks on which vehicles pass. The low frequency sound and the acceleration response of the bridge under normal traffic are measured to investigate how bridge vibrations affect the low frequency sound observed near the bridge. Observations demonstrate that strong relationships exist between frequency characteristic of bridge's acceleration response and the sound pressure level of low frequency sound. A noteworthy point is that the dynamic feature of the sound pressure level is mostly affected by dynamic feature of the span locating near the observation point.

• Speech Sciences
• /
• v.13 no.1
• /
• pp.129-139
• /
• 2006

In this paper, we propose two effective energy feature normalization methods for robust speech recognition in noisy environments. In the first method, we estimate the noise energy and remove it from the noisy speech energy. In the second method, we propose a modified algorithm for the Log-energy Dynamic Range Normalization (ERN) method. In the ERN method, the log energy of the training data in a clean environment is transformed into the log energy in noisy environments. If the minimum log energy of the test data is outside of a pre-defined range, the log energy of the test data is also transformed. Since the ERN method has several weaknesses, we propose a modified transform scheme designed to reduce the residual mismatch that it produces. In the evaluation conducted on the Aurora2.0 database, we obtained a significant performance improvement.

• Korean Institute of Interior Design Journal
• /
• no.30
• /
• pp.28-34
• /
• 2002

We need to think that space is not static but dynamic because it becomes wide and narrow, newly appears and disappears by human's behavior. Generally such movement in a space is a thing of feeling and dynamic about movement. But it is extending the experience of the subject by scientific technique's development and anticipation about the feature. The practical movement is actively introduced into architecture and interior design scope and occurs a trial about this movement. By using four elements -a form, hue, movement, light- kineticism becomes visual arts united with art and science. It recovers the art's sociality and arises participation of spectators. In the environment and art field kineticism is not simple ostentation but a current trial for human's mind and sensitivility. Kineticism is four-dimensional space considered by human's experience and is related to an observer, or experiential subject of space. Now the space except human's mind feature re-illuminates kineticism, that is, the field of the formative arts in the early part of the 20th century and gets to be 'the consensus space'.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2407-2424
• /
• 2022

Recently, Transformer has made great progress in video retrieval tasks due to its high representation capability. For the structure of a Transformer, the cascaded self-attention modules are capable of capturing long-distance feature dependencies. However, the local feature details are likely to have deteriorated. In addition, increasing the depth of the structure is likely to produce learning bias in the learned features. In this paper, an improved Transformer structure named TransDCS (Transformer with Dynamic Convolution and Shortcut) is proposed. A Multi-head Conv-Self-Attention module is introduced to model the local dependencies and improve the efficiency of local features extraction. Meanwhile, the augmented shortcuts module based on a dual identity matrix is applied to enhance the conduction of input features, and mitigate the learning bias. The proposed model is tested on MSRVTT, LSMDC and Activity-Net benchmarks, and it surpasses all previous solutions for the video-text retrieval task. For example, on the LSMDC benchmark, a gain of about 2.3% MdR and 6.1% MnR is obtained over recently proposed multimodal-based methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.653-669
• /
• 2024

Dynamic scene deblurring is a complex computer vision problem owing to its difficulty to model mathematically. In this paper, we present a novel approach for image deblurring with the help of the sharp reference image, which utilizes the reference image for high-quality and high-frequency detail results. To better utilize the clear reference image, we develop an encoder-decoder network and two novel modules are designed to guide the network for better image restoration. The proposed Reference Extraction and Aggregation Module can effectively establish the correspondence between blurry image and reference image and explore the most relevant features for better blur removal and the proposed Spatial Feature Fusion Module enables the encoder to perceive blur information at different spatial scales. In the final, the multi-scale feature maps from the encoder and cascaded Reference Extraction and Aggregation Modules are integrated into the decoder for a global fusion and representation. Extensive quantitative and qualitative experimental results from the different benchmarks show the effectiveness of our proposed method.

• Journal of Korea Society of Industrial Information Systems
• /
• v.6 no.3
• /
• pp.107-114
• /
• 2001

This paper describes Connected Digit Recognition with Considering Acoustic Feature in Korea. The recognition rate of connected digit is usually lower than word recognition. Therefore, speech feature parameter and acoustic feature are employed to make robust model for digit, and we could confirm the effect of Considering. Acoustic Feature throughout the experience of recognition. We used KLE 4 connected digit as database and 19 continuous distributed HMM as PLUs(Phoneme Like Units) using phonetical rules. For recognition experience, we have tested two cases. The first case, we used usual method like using Mel-Cepstrum and Regressive Coefficient for constructing phoneme model. The second case, we used expanded feature parameter and acoustic feature for constructing phoneme model. In both case, we employed OPDP(One Pass Dynamic Programming) and FSA(Finite State Automata) for recognition tests. When appling FSN for recognition, we applied various acoustic features. As the result, we could get 55.4% recognition rate for Mel-Cepstrum, and 67.4% for Mel-Cepstrum and Regressive Coefficient. Also, we could get 74.3% recognition rate for expanded feature parameter, and 75.4% for applying acoustic feature. Since, the case of applying acoustic feature got better result than former method, we could make certain that suggested method is effective for connected digit recognition in korean.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.9
• /
• pp.738-744
• /
• 2016

In this paper, a robust visual odometry system has been proposed and implemented in an environment with dynamic illumination. Visual odometry is based on stereo images to estimate the distance to an object. It is very difficult to realize a highly accurate and stable estimation because image quality is highly dependent on the illumination, which is a major disadvantage of visual odometry. Therefore, in order to solve the problem of low performance during the feature detection phase that is caused by illumination variations, it is suggested to determine an optimal threshold value in the image binarization and to use an adaptive threshold value for feature detection. A feature point direction and a magnitude of the motion vector that is not uniform are utilized as the features. The performance of feature detection has been improved by the RANSAC algorithm. As a result, the position of a mobile robot has been estimated using the feature points. The experimental results demonstrated that the proposed approach has superior performance against illumination variations.

• Journal of Broadcast Engineering
• /
• v.4 no.2
• /
• pp.187-196
• /
• 1999

This paper presents an efficient and automatic color and motion feature extraction algorithm for content-based MPEG-l video retrieval. Based on the proposed method. a video retrieval system is implemented. For color feature. the proposed algorithm considers dynamic color iRformation in video data, and thereby can overcome the limits of the previous key-frame based method. For motion feature, we utilize the motion vector in MPEG-l video with color information. and extract the color-motion feature. The proposed algorithm can solve the weakness of the previous location based motion feature method. Finally. the proposed method is evaluated within the implemented video retrieval system.

• Journal of KIISE:Software and Applications
• /
• v.29 no.6
• /
• pp.386-395
• /
• 2002

In this paper, we suggest dynamic genetic tree-maps(DGTM) using optimal features on recognizing data. The DGTM uses the genetic algorithm about the importance of features rarely considerable on conventional neural networks and introduces GTM(genetic tree-maps) using tree structure according of the priority of features. Hence, we propose the extended formula, DGTM(dynamic GTM) has dynamic functions to separate and merge the neuron of neural network along the similarity of features.