• Journal of information and communication convergence engineering
• /
• v.12 no.4
• /
• pp.263-270
• /
• 2014

Corner detection and feature extraction are essential aspects of computer vision problems such as object recognition and tracking. Feature detectors such as Scale Invariant Feature Transform (SIFT) yields high quality features but computationally intensive for use in real-time applications. The Features from Accelerated Segment Test (FAST) detector provides faster feature computation by extracting only corner information in recognising an object. In this paper we have analyzed the efficient object detection algorithms with respect to efficiency, quality and robustness by comparing characteristics of image detectors for corner detector and feature extractors. The simulated result shows that compared to conventional SIFT algorithm, the object recognition system based on the FAST corner detector yields increased speed and low performance degradation. The average time to find keypoints in SIFT method is about 0.116 seconds for extracting 2169 keypoints. Similarly the average time to find corner points was 0.651 seconds for detecting 1714 keypoints in FAST methods at threshold 30. Thus the FAST method detects corner points faster with better quality images for object recognition.

• Journal of Power Electronics
• /
• v.9 no.3
• /
• pp.507-515
• /
• 2009

This paper proposes an autonomous decentralized control for a parallel connected uninterruptible power supply (UPS) system based on a fast power detection method using a FPGA based hardware controller for a single phase system. Each UPS unit detects only its output voltage and current without communications signal exchange and a quasi dq transformation method is applied to detect the phase and amplitude of the output voltage and the output current for the single phase system. Fast power detection can be achieved based on a quasi dq transformation, which results in a realization of very fast transient response under rapid load change. In the proposed method, the entire control system is implemented in one FPGA chip. Complicated calculations are assigned to hardware calculation logic, and the parallel processing circuit makes it possible to realize minimized calculation time. Also, an Nios II CPU core is implemented in the same FPGA chip, and the software can be applied for non-time critical calculations. Applying this control system, an autonomous decentralized UPS system with very fast transient response is realized. Feasibility and stable operation are confirmed by means of an experimental setup with three UPSs connected in parallel. Also, rapid load change is applied and excellent performance of the system is confirmed in terms of transient response and stability.

• Journal of KIISE:Information Networking
• /
• v.35 no.4
• /
• pp.301-310
• /
• 2008

In the existing literature, the handover process reveals numerous problems manifested by high movement detection latency. FMIPv6 can reduce packet loss using a tunnel-based handover mechanism. However, this mechanism may cause performance degradation due to the out-of-sequence packets. Recently. Proxy Mobile IPv6 is proposed for network-based mobility management to reduce overhead in mobile node. PMIPv6 can decrease handover latency which related overhead in MN by using network agent. In this paper, we proposed optimized fast handover scheme called Fast Proxy Mobile IPv6 (EF-PMIPv6). The proposed EF-PMIPv6 can support fast handover using fast IAPP and ND schemes. Further, a mathematical analysis is provided to show the benefits of our scheme. In the analysis, various parameters are used to compare our scheme with the current procedures, while our approach focuses on the reduction of handover latency.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.25 no.5
• /
• pp.1027-1041
• /
• 2015

Recently, there are rapidly increasing cases of APT (Advanced Persistent Threat) attacks such as Verizon(2010), Nonghyup(2011), SK Communications(2011), and 3.20 Cyber Terror(2013), which cause leak of confidential information and tremendous damage to valuable assets without being noticed. Several anomaly detection technologies were studied to defend the APT attacks, mostly focusing on detection of obvious anomalies based on known malicious codes' signature. However, they are limited in detecting APT attacks and suffering from high false-negative detection accuracy because APT attacks consistently use zero-day vulnerabilities and have long latent period. Detecting APT attacks requires long-term analysis of data from a diverse set of sources collected over the long time, real-time analysis of the ingested data, and correlation analysis of individual attacks. However, traditional security systems lack sophisticated analytic capabilities, compute power, and agility. In this paper, we propose a Fast Data based real-time abnormal behavior detection system to overcome the traditional systems' real-time processing and analysis limitation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3904-3922
• /
• 2022

As a research hotspot, pedestrian detection has a wide range of applications in the field of computer vision in recent years. However, current pedestrian detection methods have problems such as insufficient detection accuracy and large models that are not suitable for large-scale deployment. In view of these problems mentioned above, a lightweight pedestrian detection and early warning method using a new model called you only look once (Yolov5) is proposed in this paper, which utilizing advantages of Yolov5s model to achieve accurate and fast pedestrian recognition. In addition, this paper also optimizes the loss function of the batch normalization (BN) layer. After sparsification, pruning and fine-tuning, got a lot of optimization, the size of the model on the edge of the computing power is lower equipment can be deployed. Finally, from the experimental data presented in this paper, under the training of the road pedestrian dataset that we collected and processed independently, the Yolov5s model has certain advantages in terms of precision and other indicators compared with traditional single shot multiBox detector (SSD) model and fast region-convolutional neural network (Fast R-CNN) model. After pruning and lightweight, the size of training model is greatly reduced without a significant reduction in accuracy, and the final precision reaches 87%, while the model size is reduced to 7,723 KB.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.15 no.4
• /
• pp.506-517
• /
• 2015

This paper presents a 1.248 Gb/s - 2.918 Gb/s low-power receiver MIPI-DigRF M-PHY with a fully digital frequency detection loop. MIPI-DigRF M-PHY should be operated in a very short training time which is $0.01{\mu}s$ the for HS-G2B mode. Because of this short SYNC pattern, clock and data recovery (CDR) should have extremely fast locking time. Thus, the quarter rate CDR with a fully digital frequency detection loop is proposed to implement a fast phase tracking loop. Also, a low power CDR architecture, deserializer and voltage controlled oscillator (VCO) are proposed to meet the low power requirement of MIPI-DigRF M-PHY. This chip is fabricated using a $0.11{\mu}m$ CMOS process, and the die area is $600{\mu}m{\times}250{\mu}m$. The power consumption of the receiver is 16 mW from the supply voltage of 1.1 V. The measured lock time of the CDR is less than 20 ns. The measured rms and peak jitter are $35.24ps_{p-p}$ and $4.25ps_{rms}$ respectively for HS-G2 mode.

• The KIPS Transactions:PartA
• /
• v.18A no.1
• /
• pp.19-24
• /
• 2011

As era of multi-core processors has arrived, transactional memory has been considered as an effective method to achieve easy and fast multi-threaded programming. Various hardware transactional memory systems such as UTM, VTM, FastTM, LogTM, and LogTM-SE, have been introduced in order to implement high-performance multi-core processors. Especially, LogTM-SE has provided study performance with an efficient memory management policy and a practical thread scheduling method through conflict detection based on signatures. However, increasing number of cores on a processor imposes the hardware complexity for signature processing. This causes overall performance degradation due to the heavy workload on signature comparison. In this paper, we propose a new architecture of multiple signature comparison to improve conflict detection of signature based transactional memory systems.

• Journal of IKEEE
• /
• v.18 no.1
• /
• pp.1-7
• /
• 2014

In this paper, we suggest an improved Viola-Jones detection framework for the efficient feature selection and the fast rejection method of the sub-window. Our object detector has low computational complexity because it rejects sub-windows until specific threshold. Owing to using same framework, detection performance is same with the existing Viola-Jones detector. We measure the number of average feature calculation about MIT-CMU test set. As a result of the experiment, the number of average feature calculation is reduced to 45.5% and the detection speed is improved about 58.5% compared with the previous algorithm.

• IEIE Transactions on Smart Processing and Computing
• /
• v.6 no.4
• /
• pp.269-275
• /
• 2017

Automatic vehicle detection is a very important function for autonomous vehicles. Conventional vehicle detection approaches are based on visible-light images obtained from cameras mounted on a vehicle in the daytime. However, unlike daytime, a visible-light image is generally dark at night, and the contrast is low, which makes it difficult to recognize a vehicle. As a feature point that can be used even in the low light conditions of nighttime, the rear lamp is virtually unique. However, conventional rear lamp-based detection methods seldom cope with atypical lamps, such as LED lamps, or flashing turn signals. In this paper, we detect atypical lamps by blurring the lamp area with a low pass filter (LPF) to make out the lamp shape. We also propose to detect flickering of the turn signal lamp in a manner such that the lamp area is vertically projected, and the maximum difference of two paired lamps is examined. Experimental results show that the proposed algorithm has a higher F-measure value of 0.24 than the conventional lamp pairing-based detection methods, on average. In addition, the proposed algorithm shows a fast processing time of 6.4 ms per frame, which verifies real-time performance of the proposed algorithm.

• Proceedings of the IEEK Conference
• /
• 2002.06c
• /
• pp.173-176
• /
• 2002

This paper describes an implementation of fast face detection algorithm. This algorithm can robustly detect human faces with unknown sizes and positions in complex backgrounds. This paper provides a powerful face detection algorithm using skin color segmenting. Skin Color is modeled by a Gaussian distribution in the HSI color space among different persons within the same race, Oriental. The main feature of the Algorithm is achieved face detection robust to illumination changes and a simple adaptive thresholding technique for skin color segmentation is employed to achieve robust face detection.