Following the advance of style transfer with Convolutional Neural Networks (CNNs), role styles in CNNs has drawn growing attention from a broader perspective. In this paper, we aim to fully leverage potential improve performance general vision tasks. We propose Style-based Recalibration Module (SRM), simple yet effective architectural unit, which adaptively recalibrates intermediate feature maps by exploiting their styles. SRM first extracts information each channel pooling, then estimates...

To develop an efficient deep neural network model that incorporates context from neighboring image sections to detect breast cancer on digital tomosynthesis (DBT) images.

This paper presents a new approach for avoiding double-spending attacks via the concept of recipient-oriented concepts in private blockchain networks. Together with stealth address and masternode at recipient sides, transactional privacy is assured transaction recipients become active which receive propagation unilaterally. Based on this approach, it can be shown that method suggested preventing paper. In addition, proposed possible to prevent double spending using verification time blocking...

Summary Polymer flooding is a proven technology to improve sweep efficiency, while being one of the most economical enhanced-oil-recovery (EOR) processes. Partially hydrolyzed polyacrylamide (HPAM) has been widely used for polymer flooding. As HPAM usage EOR increases, challenge produced-water management also raised because residual in produced water could increase oil content and unwanted viscosity discharging or reinjecting water. environmental standards regulations get more stringent, it...

This paper presents the experimental investigation of an optimal hot embossing process to prevent nanoscale thermal deformation microstructures replicated from electroless Ni mold fabricated by diamond tool-interfered scribing method. A polymer-based PMMA was with blazed profile: period 2.0 µm and depth 0.2 µm. The molding conditions, applying pressure (Pm), temperature (Tm) demolding (Td) were chosen as experiment parameters. In terms quality surface smoothness, profile, sharp edge,...

Abstract Polymer flooding is a proven technology to improve sweep efficiency, while being one of the most economical enhanced oil recovery (EOR) processes. Partially hydrolyzed polyacrylamide (HPAM) has been widely used for polymer flooding. As HPAM usage EOR increases, challenge produced water management also raised because residual in could increase total chemical oxygen demand and unwanted viscosity discharging or re-injecting water. environmental standards regulations get more stringent,...

A new simplified structure of bit-interleaved coded modulation (BICM) for second generation digital video broadcasting (DVB) standards, such as DVB-T2, DVB-C2, and DVB-NGH is proposed. In the BICM DVB parity column-twist interleavers are employed in order to avoid possible performance degradation due structural weaknesses scheme. this paper, a method proposed finding permutation-equivalent low-density parity-check codes that eliminate without interleaver, thereby simplifying standards. The...

Robotics has been used as an attractive tool in diverse educational fields. A variety of robotic platforms have contributed to teaching practical embedded programming engineering students at universities. However, most only support content with a low level skills and are unlikely high programming. This association negatively affects students, such incomprehension, decreased participation, dissatisfaction course quality, etc. Therefore, this paper proposed new platform relevant curricula...

There is a critical need for an accurate device orthodontists to know the magnitude of forces exerted on tooth by orthodontic brackets. Here, we propose new force measurement principle detect deformation elastic semi-sphere sensor. Specifically, aimed detail technical issues and feasibility sensor performance attached inner surface aligner or surface. Accurate tracking important optimal decision replacement cost reduction. A finite element (FE) model was developed, relationship between...

In this letter, protograph-based folded spatially coupled (FSC) LDPC codes are proposed. The new folded-type structure is obtained by folding the spatial coupling chain of a conventional (SC) protograph and interlacing nodes at staggered positions. proposed outperform SC over single multiple random-burst erasure channels. We extend construction connecting one-sided chains for higher resilience to burst FSC also compatible with windowed decoder codes.

A light switch and its spatial mapping constitute an unsolved, user interface problem introduced to the realm of Human Computer Interaction by Norman. In this paper, we introduce Previewable Switch which is a enhanced with touch sensor ability provide glimpse lights will turn on/off prior compressing switch. Such concept communicating result taking action known as feedforward [6]. Feedforward important element be considered in design it provides clear instant affordance what occur before...

A novel wafer edge quality inspection method based on analyzing curved-edge diffraction patterns is introduced in this paper. The patters, being collected through two approaches: (1) single point scanning and (2) areal imaging method, are affected by topography, geometry defects. After applying statistical feature extraction methods wavelet transform to the fringe data, corresponding features of various defect modes were further extracted. In summary, both proposed enabled characterization...

In blind detection, the user equipment attempts to decode a set of candidate groups find its own control information. False alarm (FA) refers an event in which decoding is successful even though signal not transmitted, and one several critical issues occurring detection problem. For CRC-aided polar code, with limited number CRC bits, FA occurs when check passes there no signal. this paper, we propose technique efficiently reduce events using two metrics are specially suited low rate high...

Transparent display has already become an icon of the future. It is described in fancy ways many movies; its transparency provides compelling visual openness. However, if every monitor see-through, you would find yourself facing practical problems immediately after start working with one them. When are writing e-mail, everything do will be exposed to people backside. Lack privacy inevitable drawback transparent displays because should have same images on both sides, which results lateral...

Contextualized word embeddings can lead to state-of-the-art performances in natural language understanding.Recently, a pre-trained deep contextualized text encoder such as BERT has shown its potential improving tasks including abstractive summarization.Existing approaches dialogue summarization focus on incorporating large model into task trained large-scale corpora consisting of news articles rather than dialogues multiple speakers.In this paper, we introduce self-supervised methods...

현실 세계의 정보는 다양한 모달리티로 존재한다. 이를 통합하기 위해 컴퓨터 비전 및 딥러닝 분야에서 다중모달 이미지 정합 연구가 진행되어왔다. 그러나 SIFT, ORB 등 알고리즘은 EO와 IR의 도메인 차이로 인해 좋은 결과가 도출되지 않는 문제가 있다. 또 기반 특징점 추출 모델에서는 데이터 취득의 어려움으로 기존 매우 제한된 상황이다. 따라서 본 논문에서는 이러한 문제들을 해결하기 IR 이미지와 유사한 합성 이미지를 EO 이미지로부터 생성하였고 통해 모델인 R2D2에 다중모달리티를 학습하였다. 환경에서 쌍을 취득하여 실험한 결과 다중모달에서 R2D2의 성공률은 단일모달에서 ORB의 성공률과 비교하여 4.6% 정도의 차이만 보여 정합이 성공적으로 이루어짐을 확인할 수 있었다. 데이터를 활용하여 학습이 가능함을 확인하였다.

In this paper, we propose a novel scheme for extracting JPEG (Joint Picture Experts Group) snap shot image from an HEVC (High Efficiency Video Coding) bitstream. Since both and adopt the DCT-based transform coding scheme, extract DCT information bitstream process data to be suitable compression. uses constant block size whereas variable sizes DCT. order unify size, use direct splitting merging in domain. Then, obtain coefficients of prediction frame by performing motion compensation The...

Researchers have developed compressed-domain computer vision algorithms. Despite the recent progress, demands for interesting point extraction in still exist. In this paper, we propose a corner detection DCT (Discrete Cosine Transform)-based compressed image. It partially decodes image to obtain data and then split into high precision data. The edge map is estimated by gradient patterns coefficients detect corners. Experimental results show that proposed method has about 6 pixels of average...