• Journal of the Institute of Convergence Signal Processing
• /
• v.6 no.1
• /
• pp.15-22
• /
• 2005

In this paper, we propose a face recognition system by using the CCD color image. We first get the face candidate image by using YCbCr color model and adaptive skin color information. And we use it initial curve of active contour model to extract face region. We use the Eye map and mouth map using color information for extracting facial feature from the face image. To obtain center point of Log-polar image, we use extracted facial feature from the face image. In order to obtain feature vectors, we use extracted coefficients from DCT and wavelet transform. To show the validity of the proposed method, we performed a face recognition using neural network with BP learning algorithm. Experimental results show that the proposed method is robuster with higher recogntion rate than the conventional method for the rotation and scale variant.

• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1282-1289
• /
• 2017

Since the study of deep convolutional neural network became prevalent, one of the important discoveries is that a feature map from a convolutional network can be extracted before going into the fully connected layer and can be used as a saliency map for object detection. Furthermore, the model can use features from each different layer for accurate object detection: the features from different layers can have different properties. As the model goes deeper, it has many latent skip connections and feature maps to elaborate object detection. Although there are many intermediate layers that we can use for semantic segmentation through skip connection, still the characteristics of each skip connection and the best skip connection for this task are uncertain. Therefore, in this study, we exhaustively research skip connections of state-of-the-art deep convolutional networks and investigate the characteristics of the features from each intermediate layer. In addition, this study would suggest how to use a recent deep neural network model for semantic segmentation and it would therefore become a cornerstone for later studies with the state-of-the-art network models.

• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.5
• /
• pp.277-281
• /
• 2013

Modern speaker verification systems based on support vector machines (SVMs) use Gaussian mixture model (GMM) supervectors as their input feature vectors, and the maximum a posteriori (MAP) adaptation is a conventional method for generating speaker-dependent GMMs by adapting a universal background model (UBM). MAP adaptation requires the appropriate amount of input utterance due to the number of model parameters to be estimated. On the other hand, with limited utterances, unreliable MAP adaptation can be performed, which causes adaptation noise even though the Bayesian priors used in the MAP adaptation smooth the movements between the UBM and speaker dependent GMMs. This paper proposes a sparse MAP adaptation method, which is known to perform well in the automatic speech recognition area. By introducing sparse MAP adaptation to the GMM-SVM-based speaker verification system, the adaptation noise can be mitigated effectively. The proposed method utilizes the L0 norm as a regularizer to induce sparsity. The experimental results on the TIMIT database showed that the sparse MAP-based GMM-SVM speaker verification system yields a 42.6% relative reduction in the equal error rate with few additional computations.

• Journal of Sensor Science and Technology
• /
• v.10 no.5
• /
• pp.286-291
• /
• 2001

We propose a new saliency map model for selecting an attended location in an arbitrary visual scene, which is one of the most important characteristics of human vision system. In selecting an attended location, an edge information can be considered as a feature basis to construct the saliency map. Edge filters are obtained from the independent component analysis(ICA) that is the best way to find independent edges in natural gray scenes. In order to reflect the non-uniform density in our retina, we use a multi-scaled pyramid input image instead of using an original input image. Computer simulation results show that the proposed saliency map model with multi-scale property successfully generates the plausible attended locations.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1996.10a
• /
• pp.315-318
• /
• 1996

The machine-part cell formation means the grouping of similar parts and similar machines into families in order to minimize bottleneck machines, bottleneck parts, and inter-cell part movements in cellular manufacturing systems and flexible manufacturing systems. The cell formation problem is knows as a kind of NP complete problems. This paper briefly introduces the cell-formation problem and proposes a cell formation method based on the Kohonen's self-organizing feature map which is a neural network model. It also shows some experiment results using the proposed method. The proposed method can be easily applied to the cell formation problem compared to other meta-heuristic based methods. In addition, it can be used to solve large-scale cell formation problems.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.7
• /
• pp.603-611
• /
• 2003

This paper proposes an effective feature compensation scheme based on speech model for achieving robust speech recognition. Conventional model-based method requires off-line training with noisy speech database and is not suitable for online adaptation. In the proposed scheme, we can relax the off-line training with noisy speech database by employing the parallel model combination technique for estimation of correction factors. Applying the model combination process over to the mixture model alone as opposed to entire HMM makes the online model combination possible. Exploiting the availability of noise model from off-line sources, we accomplish the online adaptation via MAP (Maximum A Posteriori) estimation. In addition, the online channel estimation procedure is induced within the proposed framework. For more efficient implementation, we propose a selective model combination which leads to reduction or the computational complexities. The representative experimental results indicate that the suggested algorithm is effective in realizing robust speech recognition under the combined adverse conditions of additive background noise and channel distortion.

• The Magazine of the IEIE
• /
• v.23 no.12
• /
• pp.115-125
• /
• 1996

This paper presents a new algorithm for the self-localization of a mobile robot using one degree perspective Invariant(Cross Ratio). Most of conventional model-based self-localization methods have some problems that data structure building, map updating and matching processes are very complex. Use of a simple cross ratio can be effective to the above problems. The algorithm is based on two basic assumptions that the ground plane is flat and two locally parallel sloe-lines are available. Also it is assumed that an environmental map is available for matching between the scene and the model. To extract an accurate steering angle for a mobile robot, we take advantage of geometric features such as vanishing points. Feature points for cross ratio are extracted robustly using a vanishing point and intersection points between two locally parallel side-lines and vertical lines. Also the local position estimation problem has been treated when feature points exist less than 4points in the viewed scene. The robustness and feasibility of our algorithms have been demonstrated through real world experiments In Indoor environments using an indoor mobile robot, KASIRI-II(KAist Simple Roving Intelligence).

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.6
• /
• pp.369-377
• /
• 2023

In the industrial manufacturing sector, quality control is pivotal for minimizing defect rates; inadequate management can result in additional costs and production delays. This study underscores the significance of detecting texture defects in manufactured goods and proposes a more precise defect detection technique. While the DFR(Deep Feature Reconstruction) model adopted an approach based on feature map amalgamation and reconstruction, it had inherent limitations. Consequently, we incorporated a new loss function using statistical methodologies, integrated a skip connection structure, and conducted parameter tuning to overcome constraints. When this enhanced model was applied to the texture category of the MVTec-AD dataset, it recorded a 2.3% higher Defect Segmentation AUC compared to previous methods, and the overall defect detection performance was improved. These findings attest to the significant contribution of the proposed method in defect detection through the reconstruction of feature map combinations.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.21 no.7
• /
• pp.75-83
• /
• 2007

IEC code based decision nile have been widely applied to detect incipient faults in power transformers. However, this method has a drawback to achieve the diagnosis with accuracy without experienced experts. In order to resolve this problem, we propose an artificial diagnosis algorithm to detect faults of power transformers using Self-Organizing Feature Map(SOM). The proposed method has two stages such as model construction and diagnostic procedure. First, faulty model is constructed by feature maps obtained by unsupervised learning for training data. And then, diagnosis is performed by compare feature map with it obtained for test data. Also the proposed method usぉms the possibility and degree of aging as well as the fault occurred in transformer by clustering and distance measure schemes. To demonstrate the validity of proposed method, various experiments are unformed and their results are presented.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.1
• /
• pp.9-16
• /
• 2002

This paper proposes a new 3D modeling technique using disparity-motion relationship and feature points. To generate the 3D model from real scene, generally, we need to compute depth of model vertices from the dense correspondence map over whole images. It takes much time and is also very difficult to get accurate depth. To improve such problems, in this paper, we only need to find the correspondence of some feature points to generate a 3D model of object without dense correspondence map. The proposed method consists of three parts, which are the extraction of object, the extraction of feature points, and the hierarchical 3D modeling using classified feature points. It has characteristics of low complexity and is effective to synthesize images with virtual view and to express the smoothness of Plain regions and the sharpness of edges.