• Proceedings of the IEEK Conference
• /
• 2000.07b
• /
• pp.803-806
• /
• 2000

The correspondence problem is known to be difficult to solve because false positives and false negatives almost always exist in real image sequences. In this paper, we propose a robust feature tracking algorithm considering incomplete trajectories such as entering and/or vanishing trajectories. We solve the correspondence problem as the optimal graph search problem, by considering false feature points and by properly reflecting motion characteristics. The proposed algorithm finds a local optimal correspondence so that the effect of false feature points can be minimized in the decision process. The time complexity of the proposed graph search algorithm is given by O(mn) in the best case and O(m$^2$n) in the worst case, where m and n are the number of feature points in two consecutive frames. The proposed algorithm can find trajectories correctly and robustly, which has been shown by experimental results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.2
• /
• pp.786-799
• /
• 2021

Road damage detection is important for road maintenance. With the development of deep learning, more and more road damage detection methods have been proposed, such as Fast R-CNN, Faster R-CNN, Mask R-CNN and RetinaNet. However, because shallow and deep layers cannot be extracted at the same time, the existing methods do not perform well in detecting objects with fewer samples. In addition, these methods cannot obtain a highly accurate detecting bounding box. This paper presents a Multi-level Feature Pyramids method based on M2det. Because the feature layer has multi-scale and multi-level architecture, the feature layer containing more information and obvious features can be extracted. Moreover, an attention mechanism is used to improve the accuracy of local boundary boxes in the dataset. Experimental results show that the proposed method is better than the current state-of-the-art methods.

• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.385-398
• /
• 2021

The complexity of deep learning models affects the real-time performance of gesture recognition, thereby limiting the application of gesture recognition algorithms in actual scenarios. Hence, a residual learning neural network based on a deep convolutional neural network is proposed. First, small convolution kernels are used to extract the local details of gesture images. Subsequently, a shallow residual structure is built to share weights, thereby avoiding gradient disappearance or gradient explosion as the network layer deepens; consequently, the difficulty of model optimisation is simplified. Additional convolutional neural networks are used to accelerate the refinement of deep abstract features based on the spatial importance of the gesture feature distribution. Finally, a fully connected cascade softmax classifier is used to complete the gesture recognition. Compared with the dense connection multiplexing feature information network, the proposed algorithm is optimised in feature multiplexing to avoid performance fluctuations caused by feature redundancy. Experimental results from the ISOGD gesture dataset and Gesture dataset prove that the proposed algorithm affords a fast convergence speed and high accuracy.

• Journal of Intelligence and Information Systems
• /
• v.7 no.1
• /
• pp.47-61
• /
• 2001

Lazy loaming methods including CBR have relative advantages in comparison with eager loaming methods such as artificial neural networks and decision trees. However, they are very sensitive to irrelevant features. In other words, when there are irrelevant features, larry learning methods have difficulty in comparing cases. Therefore, their performance can be degraded significantly. To overcome this disadvantage, feature weighting methods for lazy loaming methods have been studied. Most of the existing researches, however, were focused on global feature weighting. In this research, we propose a new local feature weighting method, which we shall call CBDFW. CBDFW stores classification performance of randomly generated feature weight vectors. Then, given a new query case, CBDFW retrieves the successful feature weight vectors and designs a feature weight vector fur the query case. In the test on credit evaluation domain, CBDFW showed better classification accuracy when compared to the results of previous researches.

• Korean Journal of Computational Design and Engineering
• /
• v.13 no.2
• /
• pp.131-138
• /
• 2008

With the advancements of the Internet and CAD data translation techniques, more CAD models are transferred from a CAD system to another through the network and interoperability is getting a common word in the CAD industry. However, when a CAD model is translated for an incompatible system into a neutral format such as STEP or IGES, its precious feature information is lost. When this feature information is lost, the advantage of feature based modeling is not valid any longer, and modification for the model is purely dependent on geometric and topological manipulations. However, the capabilities of the existing methods to modify these feature-independent models are limited as the modification involves a topological change in the model. To address this issue, we present a volumetric method to modify the solid models in neutral format. First, this method selectively decomposes the solid model to separate the portion of interest called feature volume. Next, the designer modifies the feature volume without concerning a topological change. Finally, the feature volume is united with the original solid model to complete the modification process. The results of test cases are presented to attest the usefulness of the proposed method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.4
• /
• pp.1748-1759
• /
• 2018

Weber Local Descriptor (WLD) is a stable and effective feature extraction algorithm, which is based on Weber's Law. It calculates the differential excitation information and direction information, and then integrates them to get the feature information of the image. However, WLD only considers the center pixel and its contrast with its surrounding pixels when calculating the differential excitation information. As a result, the illumination variation is relatively sensitive, and the selection of the neighbor area is rather small. This may make the whole information is divided into small pieces, thus, it is difficult to be recognized. In order to overcome this problem, this paper proposes Weber Symmetrical Local Graph Structure (WSLGS), which constructs the graph structure based on the $5{\times}5$ neighborhood. Then the information obtained is regarded as the differential excitation information. Finally, we demonstrate the effectiveness of our proposed method on the database of ORL, JAFFE and our own built database, high-definition infrared faces. The experimental results show that WSLGS provides higher recognition rate and shorter image processing time compared with traditional algorithms.

• Journal of Mechanical Science and Technology
• /
• v.18 no.8
• /
• pp.1349-1357
• /
• 2004

Computer-Aided Inspection Planning (CAIP) is the integration bridge between CAD/CAM and Computer Aided Inspection (CAI). A CAIP system for On-Machine Measurement (OMM) is proposed to inspect the complicated mechanical parts efficiently during machining or after machining. The inspection planning consists of Global Inspection Planning (GIP) and Local Inspection Planning (LIP). In the GIP, the system creates the optimal inspection sequence of the features in a part by analyzing the various feature information such as the relationship of the features, Probe Approach Directions (PAD), etc. Feature groups are formed for effective planning, and special feature groups are determined for sequencing. The integrated process and inspection plan is generated based on the sequences of the feature groups and the features in a feature group. A series of heuristic rules are developed to accomplish it. In the LIP of Part II, the system generates inspection parameters. The integrated inspection planning is able to determine optimum manufacturing sequence for inspection and machining processes. Finally, the results are simulated and analyzed to verify the effectiveness of the proposed CAIP.

• Journal of Broadcast Engineering
• /
• v.18 no.2
• /
• pp.319-322
• /
• 2013

Image retrieval research activity has moved its focus from global descriptors to local descriptors of feature point such as SIFT. MPEG is Currently working on standardization of effective coding of location and local descriptors of feature point in the context mobile based image search driven application in the name of MPEG-7 CDVS (Compact Descriptor for Visual Search). The extracted feature points consist of two parts, location information and Descriptor. For efficient image retrieval, we proposed a novel method that is gradual block-based efficient lossy location coding to compress location information according to distribution in images. From experimental result, the number of average bits per feature point reduce 5~6% and the accuracy rate keep compared to state of the art TM 3.0.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.2
• /
• pp.102-112
• /
• 2012

In this paper, we propose a texture classification using wavelet-domain BDIP (block difference of inverse probabilities) and BVLC (block variance of local correlation coefficients) features with WPCA (whitened principal component analysis) classifier. In the proposed method, the wavelet transform is first applied to a query image. The BDIP and BVLC operators are next applied to the wavelet subbands. Global moments for each subband of BDIP and BVLC are then computed and fused into a feature vector. In classification, the WPCA classifier, which is usually adopted in the face identification, searches the training feature vector most similar to the query feature vector. Experimental results show that the proposed method yields excellent texture classification with low feature dimension for test texture image DBs.

• The Journal of Korean Institute of Information Technology
• /
• v.16 no.12
• /
• pp.75-83
• /
• 2018

Recently, autonomous vehicles have been actively studied. Pedestrian detection and recognition technology is important in autonomous vehicles. Pedestrian detection using CNN(Convolutional Neural Netwrok), which is mainly used recently, generally shows good performance, but there is a performance degradation depending on the environment of the image. In this paper, we propose a pedestrian detection system applying long-term memory structure of hippocampal neural network based on CNN network with LGP-FL (Local Gradient Pattern-Feature Layer) added. First, change the input image to a size of $227{\times}227$. Then, the feature is extracted through a total of 5 layers of convolution layer. In the process, LGP-FL adds the LGP feature pattern and stores the high-frequency pattern in the long-term memory. In the detection process, it is possible to detect the pedestrian more accurately by detecting using the LGP feature pattern information robust to brightness and color change. A comparison of the existing methods and the proposed method confirmed the increase of detection rate of about 1~4%.