• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.7
• /
• pp.1726-1748
• /
• 2024

Trajectory planning is vital for autonomous systems like robotics and UAVs, as it determines optimal, safe paths considering physical limitations, environmental factors, and agent interactions. Recent advancements in trajectory planning and future location prediction stem from rapid progress in machine learning and optimization algorithms. In this paper, we proposed a novel framework for Spatial-temporal transformer-based feed-forward neural networks (STTFFNs). From the traffic flow local area point of view, skip-gram model is trained on trajectory data to generate embeddings that capture the high-level features of different trajectories. These embeddings can then be used as input to a transformer-based trajectory planning model, which can generate trajectories for new objects based on the embeddings of similar trajectories in the training data. In the next step, distant regions, we embedded feedforward network is responsible for generating the distant trajectories by taking as input a set of features that represent the object's current state and historical data. One advantage of using feedforward networks for distant trajectory planning is their ability to capture long-term dependencies in the data. In the final step of forecasting for future locations, the encoder and decoder are crucial parts of the proposed technique. Spatial destinations are encoded utilizing location-based social networks(LBSN) based on visiting semantic locations. The model has been specially trained to forecast future locations using precise longitude and latitude values. Following rigorous testing on two real-world datasets, Porto and Manhattan, it was discovered that the model outperformed a prediction accuracy of 8.7% previous state-of-the-art methods.

• Smart Media Journal
• /
• v.13 no.7
• /
• pp.9-17
• /
• 2024

Due to the increase in energy consumption, and eco-friendly policies, there is a need for efficient energy consumption in buildings. Anomaly power detection based on deep learning are being used. Because of the difficulty in collecting anomaly data, anomaly detection is performed using reconstruction error with a Recurrent Neural Network(RNN) based autoencoder. However, there are some limitations such as the long time required to fully learn temporal features and its sensitivity to noise in the train data. To overcome these limitations, this paper proposes the TCN-USAD, combined with Temporal Convolution Network(TCN) and UnSupervised Anomaly Detection for multivariate data(USAD). The proposed model using TCN-based autoencoder and the USAD structure, which uses two decoders and adversarial training, to quickly learn temporal features and enable robust anomaly detection. To validate the performance of TCN-USAD, comparative experiments were performed using two building energy datasets. The results showed that the TCN-based autoencoder can perform faster and better reconstruction than RNN-based autoencoder. Furthermore, TCN-USAD achieved 20% improved F1-Score over other anomaly detection models, demonstrating excellent anomaly detection performance.

• Applied Microscopy
• /
• v.51
• /
• pp.5.1-5.7
• /
• 2021

The olfactory organs of two African lungfishes, Protopterus amphibius and P. dolloi, were investigated using a stereo microscope and a compound light microscope and were described anatomically, histologically, and histochemically. Like other lungfishes, these species present the following general features: i) elongated olfactory chamber (OC), ii) anterior nostril at the ventral tip of the upper lip, iii) posterior nostril on the palate of the oral cavity, iv) lamellae with multiple cell types such as olfactory receptor neurons, supporting cells, basal cells, lymphatic cells, and mucous cells (MC), and vi) vomero-like epithelial crypt (VEC) made of glandular epithelium (GE) and crypt sensory epithelium. Some of these features exhibit differences between species: MCs are abundant in both the lamellar and inner walls of the OC in P. amphibius but occur only in lamellae in P. dolloi. On the other hand, some between feature differences are consistent across species: the GE of both P. amphibius and P. dolloi is strongly positive for Alcian blue (pH 2.5)-periodic acid Schiff (deep violet coloration), and positive with hematoxylin and eosin and with Masson's trichrome (reddish-brown staining), unlike the MCs of the two species which stain dark red with both Alcian blue (pH 2.5)-periodic acid Schiff and Masson's trichrome but respond faintly to hematoxylin and eosin. The differing abundance of MCs in the two lungfishes might reflect different degrees in aerial exposure of the olfactory organ, while the neutral and acid mucopolysaccharide-containing VEC, as indicated by staining properties of the MCs, is evolutionary evidence that P. amphibius and P. dolloi are the closest living relatives to tetrapods, at least in the order Dipnoi.

• Journal of the Semiconductor & Display Technology
• /
• v.23 no.3
• /
• pp.127-134
• /
• 2024

Products were typically produced using specialized equipment such as CNC machines, milling machines, and lathes in traditional manufacturing. However, modern manufacturing is increasingly attempting with technological advancements to leverage large industrial robots for machining, offering greater flexibility, efficiency, and a high degree of freedom throughout the entire production process. Additionally, the demand for industrial robots continues to rise as industries adopt smart factories. These robots are becoming larger, more precise, and faster, as they take over tasks previously requiring specialized equipment or skilled human operators. Where numerous robots are in operation in factories, ensuring a stable supply chain and maintaining operational uptime is crucial. Therefore, preparing for potential mechanical failures in each robot is necessary, and there is a growing need for technologies that enable real-time fault diagnosis and predictive maintenance. A large industrial robot used for machining was employed as a testbed for fault diagnosis in this study. The Vibration data was collected from various robot axes under both normal operating conditions and abnormal conditions, such as end-effector overloads and drive malfunctions. The collected vibration data was then preprocessed, and key features were analyzed and extracted. The extracted features were used to build a learning model, and in this study, the CNN (Convolutional Neural Network) algorithm was applied instead of k-NN (k-Nearest Neighbors) to diagnose defects occurring in the discontinuous movements of the robot, thereby improving accuracy.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.

• Journal of the Korean earth science society
• /
• v.21 no.3
• /
• pp.276-286
• /
• 2000

A 2.3 m-long core obtained from Marian Cove, King George Island in the South Shetland Islands, West Antarctica shows clues to the glaciomarine sedimentation during the Holocene. The lower part below 115cm-deep of the core is predominated by coarser material such as diamictons compared with the higher part above 105cm dominated by finer material (rhythmite and massive muds). Based on the granulometric features the coarse materials are generally supposed to be glacially-driven and basal tills, whereas the fine materials appear to originate from various sources such as meltwater-supplied, glacially-supplied, wind-blown, and organic origins. However, the presence of erratic coarse particles in the finer materials suggests the ice-rafted origin of the relevant materials. The lower part below 105cm-deep of the core was characterized by lower TN, TC, and TOC contents, and by higher TS and CaCO$_3$ contents compared with its upper part. No significant changes in C/N ratio were shown throughout the core. The ice cliff along the east side of Marian Cove seemed to locate to the west by 1.6km at 8,300 years B. P. on the basis of the repetitive occurrence of rhythmite and diamicton. Since the retreat of ice cliff in 7,970${\pm}$70 years B. P. the sediments of Marian Cove were dominated by fine materials and ice-rafted materials. The abrupt increase of coarse materials in 175cm-4 deep seems to result from supply of coarse materials due to earthquake or other drastic phenomena.

• Journal of the Korean earth science society
• /
• v.21 no.5
• /
• pp.635-646
• /
• 2000

Geological achievements of the 20th century revolutionized our views about geological understanding and concept. A good example is the concept of continental drift suggested early in the 20th century and later explained in terms of seafloor spreading and plate tectonics. Our understanding of the compositions of materials forming earth has also improved during the20th century. Radio and stable isotopes together with biostratigraphy and sequence stratigraphy allow us to interpret the evolution of sedimentary basins in terms of plate movement and sedimentation processes. The Deep Sea Drilling Project initiated in 1960s and continued as the Ocean Drilling Project in 1980s is one of the most successful international research observations, and new developments in computational techniques have provided a wholly new view about the interior of the earth. Most of the geological features and phenomena observed in deep sea and around continental margins are now explained in terms of global tectonic processes such as superplumes flowing up from the interior of our planet and interacting with such as Rodinia Pannotia and Nena back in the Precambrian time. The space explorations which began in the late 1950s opened up a new path to astrogeology, astrobiology, and astropaleontology. The impact theory rooted in the discovery of iridium and associated phenomena in 1980s revived Cuvier's catastrophism as a possible explanation for the extinctions of biotas found in the geological record of this planet. Due to the geological achievements made in the 20th century, we now have a better understanding of geologic times and processes that were too long to be grasped by human records.

• Korean Journal of Remote Sensing
• /
• v.36 no.3
• /
• pp.449-460
• /
• 2020

Super-resolution is a technique used to reconstruct an image with low-resolution into that of high-resolution. Recently, deep-learning based super resolution has become the mainstream, and applications of these methods are widely used in the remote sensing field. In this paper, we propose a super-resolution method based on the deep back-projection network model to improve the satellite image resolution by the factor of four. In the process, we customized the loss function with the edge loss to result in a more detailed feature of the boundary of each object and to improve the stability of the model training using generative adversarial network based on Wasserstein distance loss. Also, we have applied the detail preserving image down-scaling method to enhance the naturalness of the training output. Finally, by including the modified-residual learning with a panchromatic feature in the final step of the training process. Our proposed method is able to reconstruct fine features and high frequency information. Comparing the results of our method with that of the others, we propose that the super-resolution method improves the sharpness and the clarity of WorldView-3 and KOMPSAT-2 images.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.5
• /
• pp.423-431
• /
• 2016

In general LPR(License Plate Recognition) in outdoor image is not so simple differently from in the image captured from manmade environment, because of geometric shape distortion and large illumination changes. this paper proposes three techniques for LPR in outdoor images captured from CCTV. At first, a serially connected multi-stage Adaboost LP detector is proposed, in which different complementary features are used. In the proposed detector the performance is increased by the Haar-like Adaboost LP detector consecutively connected to the MB-LBP based one in serial manner. In addition the technique is proposed that makes image processing easy by the prior determination of LP type, after correction of geometric distortion of LP image. The technique is more efficient than the processing the whole LP image without knowledge of LP type in that we can take the appropriate color to gray conversion, accurate location for separation of text/numeric character sub-images, and proper parameter selection for image processing. In the proposed technique we use DBN(Deep Belief Network) to achieve a robust character recognition against stroke loss and geometric distortion like slant due to the incomplete image processing.

• Korean Journal of Materials Research
• /
• v.11 no.10
• /
• pp.889-894
• /
• 2001

2wt% $Al_2O_3-doped$ ZnO (AZO) thin films were deposited on sapphire (0001) single crystal substrate by parellel type rf magnetron sputtering at 55$0^{\circ}C$. The as-grown AZO thin films was polycrystalline and showed only broad deep defect-level photoluminescence (PL). In order to examine the change of PL property, AZO thin films were annealed in $N_2$ (N-AZO) and $H_2$ (H-AZO) at the temperature of $600^{\circ}C$~$1000^{\circ}C$ through rapid thermal annealing. After annealed at $800^{\circ}C$, N-AZO shows near band edge emission (NBE) with very small deep-level emission, and then N-AZO annealed at $900^{\circ}C$ shows only sharp NBE with 219 meV FWHM. In Comparison with N-AZO, H-AZO exhibits very interesting PL features. After $600^{\circ}C$ annealing, deep defect-level emission was quire quenched and NBE around 382 nm (3.2 eV) was observed, which can be explained by the $H_2$passivation effect. At elevated temperature, two interesting peaks corresponding to violet (406 nm, 3.05 eV) and blue (436 nm, 2.84 eV) emission was firstly observed in AZO thin films. Moreover, peculiar PL peak around 694 nm (1.78 eV) is also firstly observed in all the H-AZO thin films and this is believed good evidence of hydrogenation of AZO. Based on defect-level scheme calculated by using the full potential linear muffin-tin orbital (FP-LMTO), the emission 3.2 eV, 3.05 eV, 3.84 eV and 1.78 eV of H-AZO are substantially deginated as exciton emission, transition from conduction band maximum to $V_{ Zn},$ from $Zn_i$, to valence band maximum $(V_{BM})$ and from $V_{o} to V_BM}$, respectively.