• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.7 no.1
• /
• pp.32-41
• /
• 2008

The moving object tracking in vision based observation using video uses difference method between GMM(Gaussian Mixture Model) based background and present image. In the case of racking object using binary image made by threshold, the object is merged not by object information but by Cast-Shadow. This paper proposed the method that eliminates Cast-Shadow using backpropagation Neural Network. The neural network is trained by abstracting feature value form training image of object range in 10-movies and Cast-Shadow range. The method eliminating Cast-Shadow is based on the method distinguishing shadow from binary image, its Performance is better(16.2%, 38.2%, 28.1%, 22.3%, 44.4%) than existing Cast-Shadow elimination algorithm(SNP, SP, DNM1, DNM2, CNCC).

• Journal of the Korea Society of Computer and Information
• /
• v.10 no.4 s.36
• /
• pp.17-27
• /
• 2005

In this study, it proposes the neural network system for character recognition and restoration. Proposes system composed by recognition part and restoration part. In the recognition part. it proposes model of effective pattern recognition to improve ART Neural Network's performance by restricting the unnecessary top-down frame generation and transition. Also the location feature extraction algorithm which applies with Hangeul's structural feature can apply the recognition. In the restoration part, it composes model of inputted image's restoration by Hopfield neural network. We make part experiments to check system's performance, respectively. As a result of experiment, we see improve of recognition rate and possibility of restoration.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.3
• /
• pp.694-701
• /
• 2004

In this paper, we proposed a feature-based tracking system that traces moving objects with a pan-tilt camera after separating the global motion of an active camera and the local motion of moving objects. The tracking system traces only the local motion of the comer features in the foreground objects by finding the block motions between two consecutive frames using a block-based motion estimation and eliminating the global motion from the block motions. For the robust estimation of the camera motion using only the background motion, we suggest a dominant motion extraction to classify the background motions from the block motions. We also propose an efficient clustering algorithm based on the attributes of motion trajectories of corner features to remove the motions of noise objects from the separated local motion. The proposed tracking system has demonstrated good performance for several test video sequences.

• Journal of Mechanical Science and Technology
• /
• v.19 no.12
• /
• pp.2213-2223
• /
• 2005

In order to reconstruct a full 3D human model in reverse engineering (RE), a 3D scanner needs to be placed arbitrarily around the target model to capture all part of the scanned surface. Then, acquired multiple scans must be registered and merged since each scanned data set taken from different position is just given in its own local co-ordinate system. The goal of the registration is to create a single model by aligning all individual scans. It usually consists of two sub-steps: rough and fine registration. The fine registration process can only be performed after an initial position is approximated through the rough registration. Hence an automated rough registration process is crucial to realize a completely automatic RE system. In this paper an automated rough registration method for aligning multiple scans of complex human face is presented. The proposed method automatically aligns the meshes of different scans with the information of features that are extracted from the estimated principal curvatures of triangular meshes of the human face. Then the roughly aligned scanned data sets are further precisely enhanced with a fine registration step with the recently popular Iterative Closest Point (ICP) algorithm. Some typical examples are presented and discussed to validate the proposed system.

• Journal of Broadcast Engineering
• /
• v.24 no.6
• /
• pp.939-947
• /
• 2019

With the revitalization of the media market, the necessity of compression, searching, editing and copyright protection of videos is increasing. In this paper, we propose a method to detect scene change in all these fields. We propose a pre-processing, feature point extraction using SIFT, and matching algorithm for detecting the same scene change even if distortions such as resolution change, subtitle insertion, compression, and flip are added in the distribution process. Also, it is applied to filtering technology and it is confirmed that it is effective for all transformations other than considering transform.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2006.11a
• /
• pp.203-207
• /
• 2006

Nowadays, Cars are continuing to grow at an alarming rate but they also cause many problems such as traffic accident, pollutions and so on. One of the most effective methods that prevent traffic accidents is the use of traffic monitoring systems, which are already widely used in many countries. The monitoring system is beginning to be used in domestic recently. An intelligent monitoring system generates photo images of cars as well as identifies cars by recognizing their plates. That is, the system automatically recognizes characters of vehicle plates. An automatic vehicle plate recognition consists of two main module: a vehicle plate locating module and a vehicle plate number identification module. We study for a vehicle plate number identification module in this paper. We use image preprocessing, feature extraction, multi-layer neural networks for recognizing characters of vehicle plates and we present a feature-comparison method for improving the performance of vehicle plate number identification module. In the experiment on identifying vehicle plate number, 300 images taken from various scenes were used. Of which, 8 images have been failed to identify vehicle plate number and the overall rate of success for our vehicle plate recognition algorithm is 98%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.147-165
• /
• 2021

In this work, we empirically evaluated the efficiency of the recent EfficientNetB0 model to identify and diagnose malaria parasite infections in blood smears. The dataset used was collected and classified by relevant experts from the Lister Hill National Centre for Biomedical Communications (LHNCBC). We prepared our samples with minimal image transformations as opposed to others, as we focused more on the feature extraction capability of the EfficientNetB0 baseline model. We applied transfer learning to increase the initial feature sets and reduced the training time to train our model. We then fine-tuned it to work with our proposed layers and re-trained the entire model to learn from our prepared dataset. The highest overall accuracy attained from our evaluated results was 94.70% from fifty epochs and followed by 94.68% within just ten. Additional visualization and analysis using the Gradient-weighted Class Activation Mapping (Grad-CAM) algorithm visualized how effectively our fine-tuned EfficientNetB0 detected infections better than other recent state-of-the-art DCNN models. This study, therefore, concludes that when fine-tuned, the recent EfficientNetB0 will generate highly accurate deep learning solutions for the identification of malaria parasites in blood smears without the need for stringent pre-processing, optimization, or data augmentation of images.

• Journal of Korea Multimedia Society
• /
• v.24 no.11
• /
• pp.1552-1559
• /
• 2021

Convolution neural networks (CNNs) show notable performance in image processing and are used as representative core models. CNNs extract and learn features from large amounts of train dataset. In general, it has a structure in which a convolution layer and a fully connected layer are stacked. The core of CNN is the convolution layer. The size of the kernel used for feature extraction and the number that affect the depth of the feature map determine the amount of weight parameters of the CNN that can be learned. These parameters are the main causes of increasing the computational complexity and memory usage of the entire neural network. The most computationally expensive components in CNNs are fully connected and spatial convolution computations. In this paper, we propose a Fourier Convolution Neural Network that performs the operation of the convolution layer in the Fourier domain. We work on modifying and improving the amount of computation by applying the fast fourier transform method. Using the MNIST dataset, the performance was similar to that of the general CNN in terms of accuracy. In terms of operation speed, 7.2% faster operation speed was achieved. An average of 19% faster speed was achieved in experiments using 1024x1024 images and various sizes of kernels.

• Current Optics and Photonics
• /
• v.8 no.5
• /
• pp.463-471
• /
• 2024

Multimode fibers (MMFs) possess high information throughput and small core diameter, making them highly promising for applications such as endoscopy and communication. However, modal dispersion hinders the direct use of MMFs for image transmission. By training neural networks on time-series waveforms collected from MMFs it is possible to reconstruct images, transforming blurred speckle patterns into recognizable images. This paper proposes a fully convolutional neural-network model, MSMDFNet, for image restoration in MMFs. The network employs an encoder-decoder architecture, integrating multiscale convolutional modules in the decoding layers to enhance the receptive field for feature extraction. Additionally, attention mechanisms are incorporated from both spatial and channel dimensions, to improve the network's feature-perception capabilities. The algorithm demonstrates excellent performance on MNIST and Fashion-MNIST datasets collected through MMFs, showing significant improvements in various metrics such as SSIM.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3806-3825
• /
• 2016

This paper brings out a neoteric frame of reference for visual semantic based 3d video search and retrieval applications. Newfangled 3D retrieval application spotlight on shape analysis like object matching, classification and retrieval not only sticking up entirely with video retrieval. In this ambit, we delve into 3D-CBVR (Content Based Video Retrieval) concept for the first time. For this purpose we intent to hitch on BOVW and Mapreduce in 3D framework. Here, we tried to coalesce shape, color and texture for feature extraction. For this purpose, we have used combination of geometric & topological features for shape and 3D co-occurrence matrix for color and texture. After thriving extraction of local descriptors, TB-PCT (Threshold Based- Predictive Clustering Tree) algorithm is used to generate visual codebook. Further, matching is performed using soft weighting scheme with L₂ distance function. As a final step, retrieved results are ranked according to the Index value and produce results .In order to handle prodigious amount of data and Efficacious retrieval, we have incorporated HDFS in our Intellection. Using 3D video dataset, we fiture the performance of our proposed system which can pan out that the proposed work gives meticulous result and also reduce the time intricacy.