MOTIVATION: Predicting protein structures with high accuracy is a critical challenge for the broad community of life sciences and industry. Despite progress made by deep neural networks like AlphaFold2, there is a need for further improvements in the quality of detailed structures, such as side-chains, along with protein backbone structures. RESULTS: Building upon the successes of AlphaFold2, the modifications we made include changing the losses of side-chain torsion angles and frame aligned point error, adding loss functions for side chain confidence and secondary structure prediction, and replacing template feature generation with a new alignment method based on conditional random fields. We also performed re-optimization by conformational space annealing using a molecular mechanics energy function which integrates the potential energies obtained from distogram and side-chain prediction. In the CASP15 blind test for single protein and domain modeling (109 domains), DeepFold ranked fourth among 132 groups with improvements in the details of the structure in terms of backbone, side-chain, and Molprobity. In terms of protein backbone accuracy, DeepFold achieved a median GDT-TS score of 88.64 compared with 85.88 of AlphaFold2. For TBM-easy/hard targets, DeepFold ranked at the top based on Z-scores for GDT-TS. This shows its practical value to the structural biology community, which demands highly accurate structures. In addition, a thorough analysis of 55 domains from 39 targets with publicly available structures indicates that DeepFold shows superior side-chain accuracy and Molprobity scores among the top-performing groups. AVAILABILITY AND IMPLEMENTATION: DeepFold tools are open-source software available at https://github.com/newtonjoo/deepfold.

This paper presents a mobile AR system for visualizing airflow and temperature change made by virtual air conditioners. Even though there have been efforts to integrate the results of airflow/temperature simulation into the real world via AR, they support neither interactive modeling of the environments nor real-time simulation. This paper presents an AR system, where 3D mapping and air conditioner installation are made interactively, and then airflow/temperature simulation and visualization are made at real time. The proposed system is designed in a client-server architecture, where the server is in charge of simulation and the rest is taken by the client.

The inspection of sewing defects is an essential step in the quality assurance of garment manufacturing. Although traditional automated defect detection applications have shown good performance, these methods are usually configured with handcrafted features designed by a human operator. Recently, deep learning methods that include Convolutional Neural Networks (CNNs) have demonstrated excellent performance in a wide variety of computer-vision applications. To take advantage of the CNN’s feature representation, the direct utilization of feature maps from the convolutional layers as universal feature descriptors has been studied. In this paper, we propose a sewing defect detection method using a CNN feature map extracted from the initial layers of a pre-trained VGG-16 to detect a broken stitch from a captured image of a sewing operation. To assess the effectiveness of the proposed method, experiments were conducted on a set of sewing images, including normal images, their synthetic defects, and rotated images. As a result, the proposed method detected true defects with 92.3% accuracy. Moreover, additional conditions for computing devices and deep learning libraries were investigated to reduce the computing time required for real-time computation. Using a general and cheap single-board computer with resizing the image and utilizing a lightweight deep learning library, the computing time was 0.22 s. The results confirm the feasibility of the proposed method’s performance as an appropriate manufacturing technology for garment production.

We propose a novel integrated framework to combine the self-learning super-resolution (SR) with dual-learning noise-reduction (NR) for compressed images. Contrary to existing learning based denoising approach, dual-learning based joint SR and NR is proposed by adding a denoised training set. It makes the proposed framework more suitable for highly compressed noise by referring to closer patch in a training set. Also, it is robust for SR artifacts since the joint framework is designed in such a way that one could learn a process to simultaneously perform NR and SR. Experimental results show that the proposed joint SR and NR framework can achieve higher objective and subjective qualities, compared with individual processing of NR and SR.

Keyword-based search exploits the exact match between the index terms of a query and documents. Thus, some documents, although they are relevant to the given query, may not be returned to users unless the documents include the index terms of the query. Some search engines use the authority of documents, which is derived from the links of documents, to help keyword-based search provide more accurate search results. However, unlike the Web documents, if the links between documents do not exist, it is difficult to exploit the authority for ranking documents. In this paper, our goals are to derive the implicit authority of documents that do not have explicit links through semantic collaborative filtering (SCF), and to retrieve documents that are semantically related to the given query. To achieve these goals, we represent users' preferences, queries and documents with their corresponding concepts by extending a Bayesian belief network. It is because the Bayesian belief network provides a clear formalism for mapping the users' preferences, queries and documents to their corresponding concepts. The concepts are extracted from a taxonomic knowledgebase such as the Open Directory Project Web directory. In our experiment, we have shown that the extended Bayesian belief network using taxonomic knowledge outperforms the conventional approaches for personalized search.

We present a Korean question answering framework for restricted domains, called K-QARD. K-QARD is developed to achieve domain portability and robustness, and the framework is successfully applied to build question answering systems for several domains.

Pitch tracking and pitch marking (PM) are two important speech signal analysis techniques for several applications. The accuracy of both pitch marking and tracking is significant to generate smooth synthesized speech by controlling the pitch and duration of voiced speech in Text-to-Speech (TTS) system for example. In this paper, we present a novel hybrid approach, combining electro-glotto-graph (EGG)-based PM and speech signal-based PM into a single framework, to acquire more reliable and automatic PM technique. Experimental results show that the PM performance of the suggested method is excellent being capable of determining Glottal Closure Instants (GCI) precisely even in the case of noisy EGG signals.

In this paper, we describe the effect of actuator relocation on singularity, Jacobian and kinematic isotropy performance of parallel robots and manipulators. The actuator relocation is related to the location where the actuator is mounted, and the actuated joint and transmission ratio of actuators with gears. The actuators can be located far from the joint they control if an appropriate transmission such as a pulley, gear and linkage is used, which is general in industrial serial robots. Since parallel robots consist of several serial legs, it is necessary to study the effect on the J/sub q/ and J/sub x/ when the actuator in each leg of the parallel robot is mounted remotely and the actuation joint is driven through a transmission. First, we describe the effect of both transmission ratios and actuator location on Jacobian isotropy and singularity. We illustrate the effect of actuator relocation on kinematic isotropy performance with examples. This actuator relocation can be used to optimize the kinematic performance of parallel robots.

We propose a geometric data fusion (GDF) method using a perception-net which can provide error reducing, uncertainty management, and maintaining consistency. We propose a perception-net to design a state estimator for dynamic systems and apply the proposed geometric data fusion method to obtain the optimal estimate, propagate uncertainties and utilize the system knowledge. We present comparisons between the proposed estimator and the conventional estimators. It is also shown that the additional priori information on the system can be easily utilized in the proposed estimator to improve the performance. Through illustrative examples, it is verified that the proposed estimator presents better performances than existing filters and improves performances via utilizing system knowledge.

Describes two decision making strategies of the lane departure warning system for a vehicle and provides experimental results to assess the performance of each strategy. The overview of the lane departure warning system mounted on the vehicle is presented. A lateral offset (LO) based strategy and a time to lane crossing (TLC) based strategy are proposed to be used in the warning module which detects unintended lane departure so as to warn the driver. The performance criteria of the warning strategies are defined as: 1) false alarm rate, 2) alarm triggering time. The proposed strategies are incorporated in our lane departure warning system to be tested in real expressway experiments. The experimental results for the performance optimization of both strategies are discussed and compared with each other.

Proposed is an algorithm to estimate vehicle pose and road curvature by geometrically fusing sensor data from camera image, velocity meter, and steering wheel encoder. To achieve computational efficiency in processing in a real time sequence, we propose a method to model the lane on the road as a series of connected rectangular plates. We propose an algorithm, the so called "Perception-Net", where not only variables denoting the vehicle pose and the road curvature, but also the corresponding uncertainties are propagated in forward and backward directions in such a way to satisfy the given constraint condition, maintain consistency, reduce the uncertainties, and guarantee robustness. An experimental result is also presented.

Proposes a vision based lane departure warning system which warns the driver of drifting off the lane inattentively. An estimation algorithm is proposed to obtain the vehicle pose and lane geometry data so as to detect an unintended lane departure. The lane departure alarm is triggered by a warning algorithm which makes decisions on giving an alarm based on the data from the estimation algorithm. The performance criteria of the lane departure warning system are defined as false alarm rate and alarm triggering time. The system is implemented in a test automobile and tested in real expressway experiments optimizing the performance. The experimental results demonstrate the applicability of the proposed system.