• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.9
• /
• pp.349-358
• /
• 2021

Recently, a method of finding the attention area or localization area for an object of an image using CAM (Class Activation Map)[1] has been variously carried out as a study of WSOL (Weakly Supervised Object Localization). The attention area extraction from the object heat map using CAM has a disadvantage in that it cannot find the entire area of the object by focusing mainly on the part where the features are most concentrated in the object. To improve this, using CAM and Selective Search[6] together, we first expand the attention area in the heat map, and a Gaussian smoothing is applied to the extended area to generate retraining data. Finally we train the data to expand the attention area of the objects. The proposed method requires retraining only once, and the search time to find an localization area is greatly reduced since the selective search is not needed in this stage. Through the experiment, the attention area was expanded from the existing CAM heat maps, and in the calculation of IOU (Intersection of Union) with the ground truth for the bounding box of the expanded attention area, about 58% was improved compared to the existing CAM.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.10
• /
• pp.955-961
• /
• 2006

The global ultrasonic system for the self-localization of a mobile robot consists of several ultrasonic transmitters fixed at some reference positions in the global coordinates of robot environment. By activating the ultrasonic transmitters, the mobile robot is able to get the distance to the ultrasonic transmitters and compute its own position in the global coordinate. Due to the limitation on the ultrasonic signal strength and beam width as well as the environmental obstacles however, the ultrasonic signals from some generator may not be transmitted to the robot. Thus, instead of activating the all ultrasonic transmitters, it is necessary to select some ultrasonic generators to activate based on the current robot position. In this paper, we propose a selective activation algorithm for self-localization with the global ultrasonic system. The selective activation algorithm gets the meaningful ultrasonic data at every sampling instants, which results in the faster and more accurate response of the self-localization than the conventional sequential activation. Through the self-localization and path following control, we verify the effectiveness of the proposed selective activation algorithm.

• Journal of Advanced Marine Engineering and Technology
• /
• v.37 no.8
• /
• pp.932-938
• /
• 2013

Recently, the Fingerprint, which is one of the methods of indoor localization using WLAN, has been many studied owing to robustness about ranging error by the diffraction and refraction of radio waves. However, in the signal gathering process and comparison operation for the measured signals with the database, this method requires time consumption and computational complexity. In order to compensate for these problems, this paper presents, based on proposed modified database, WLAN indoor localization algorithm using selective operation of collected signal in real time. The proposed algorithm reduces the configuration time and the size of the data in the database through linear interpolation and thresholding according to the signal strength, the localization accuracy, while reducing the computational complexity, is maintained through selective operation of the signals which are measured in real time. The experimental results show that the accuracy of localization is improved to 17.8% and the computational complexity reduced to 46% compared to conventional Fingerprint in the corridor by using proposed algorithm.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.3 no.2
• /
• pp.109-118
• /
• 2008

We present a robust localization algorithm using particle filter for mobile robots in a dynamic environment. It is difficult to describe moving obstacles like people or other robots on the map and the environment is changed after mapping. A mobile robot cannot estimate its pose robustly with this incomplete map because sensor observations are corrupted by un-modeled obstacles. The proposed algorithms provide robustness in such a dynamic environment by suppressing the effect of corrupted sensor observations with a selective update or a sampling from non-corrupted window. A selective update method makes some particles keep track of the robot, not affected by the corrupted observation. In a sampling from non-corrupted window method, particles are always sampled from several particle sets which use only non-corrupted observation. The robustness of proposed algorithm is validated with experiments and simulations.

• The Journal of Korea Robotics Society
• /
• v.6 no.1
• /
• pp.58-68
• /
• 2011

This paper proposes a technique of indoor localization for mobile robot by so called indoor GPS and EKF. Basically the concept of indoor GPS is similar outdoor GPS, and the indoor GPS gets distances between Anchors and Tag by a ranging method of CSS and then estimates the coordinate by distances and known Anchor positions. After we performed performance test of indoor GPS system in ideal and multipath environment, we configured that the indoor GPS has internal error factors and external error factors. This paper handled a multipath problem belonging to external error factors. At first various direct physical method are introduced to fix the multipath problems, and as expected we got errors corrected considerably. And then the method of selective anchors for indoor GPS is applied. With these two level improvement of indoor GPS performance, EKF(Extended Kalman Filter) is applied to mobile robot in indoor environment. The usefulness of the proposed methods are shown by a series of experiments in a environment giving contaminated data by multipath.

• Journal of Gastric Cancer
• /
• v.17 no.3
• /
• pp.220-227
• /
• 2017

Purpose: Precise localization of tumors and creation of sufficient proximal resection margins are complicated processes during total laparoscopic distal gastrectomy (TLDG) for clinical T1/T2 gastric cancers. Various solutions to this problem have also yielded many disadvantages. In this study, we reviewed a preoperative endoscopic clipping method based on the results of negative biopsy and selective intraoperative radiography. Materials and Methods: A retrospective review of 345 consecutive patients who underwent TLDG and preoperative endoscopic clipping for tumor localization was conducted. During preoperative endoscopy, the endoscopists performed negative biopsies just 1-2 cm selectively above the tumor's upper limit. After confirming the biopsy results, endoscopic metal clips were applied just proximal to the negative biopsy site the day before surgery. Selective intraoperative tumor localization using portable abdominal radiography was performed only when we could not ensure a precise resection line. Results: Negative biopsy was performed in 244 patients. Larger tumor size (P=0.008) and more distally located tumors (P=0.052) were observed more frequently in the negative biopsy group than in the non-negative biopsy group. The non-negative biopsy group had significantly higher frequencies of differentiated tumor types than the negative biopsy group (P=0.003). Of the 244 patients who underwent negative biopsies, 6 had cancer cells in their biopsy specimens. We performed intraoperative radiography in 12 patients whose tumors had difficult-to-determine proximal margins. No tumors were found in the proximal resection margins of any patients. Conclusions: Our tumor localization method is a promising and accurate method for securing a sufficient resection margin during TLDG.

• Clean Technology
• /
• v.24 no.2
• /
• pp.105-111
• /
• 2018

This study presents a novel and eco-friendly process that can precisely control the location of the patches on the patch particles. The method of manufacturing these anisotropic hexagram patch particles consists of sequential combinations of two separate methods such as a sequential micromolding technique for fabricating patch particles and a selective localization method for controlling the location of patches on the patch particles. The micromolding technique was carried out using physicochemically stable material as a micromold. In order to fabricate the highly stable patch anisotropic hexagram particles, the perfluoropolyether (PFPE) micromold was used to the process of the micromolding technique because they could prevent the problem of diffusion of hydrophobic monomers while conventional poly(dimethylsiloxane) (PDMS) micromold is limited to prevent the problem of diffusion of hydrophobic monomers. Based on combination methods of the micromolding technique and the selective localization method, the reproducibility and stability have been improved to fabricate 12 different types of anisotropic hexagram patch particles. This fabrication method shows the unique advantages in eco-friend condition, easy and fast fabrication due to less number of process, the feasibility of a mass production. We believe that these anisotropic hexagram patch particles can be widely utilized to the field of the directional self-assembly.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.2
• /
• pp.116-124
• /
• 2011

A service robot can identify its own position relative to landmarks, the locations of which are known in advance. The main contribution of this research is that it gives various ways of making the self-localizing error smaller by referring to special landmarks which are developed as high gain reflection material and coded array associations. In this paper, the authors propose a set of indices to evaluate the accuracy of self-localizing methods using the selective reflection landmark and infrared projector, and the indices are derived from the sensitivity enhancement using 3D distortion calibration of camera. And then, the accurarcy of self-localizing a mobile robot with landmarks based on the indices is evaluated, and a rational way to minimize to reduce the computational cost of selecting the best self-localizing method. The simulation results show a high accuracy and a good performance.

• Journal of Life Science
• /
• v.25 no.8
• /
• pp.947-952
• /
• 2015

Prostate apoptosis response-4 (Par-4) was originally identified in androgen-independent prostate cancer cells undergoing apoptosis. Par-4 is ubiquitously expressed in normal cells and tissues, but it is downregulated in several types of cancers. Par-4 is a 38 kDa tumor suppressor protein encoded by the PARW gene. Par-4 promotes apoptosis in a variety of cancerous cells, but not in normal cells. In this review, we focused on the structure, expression and function of Par-4 in apoptotic signaling pathway. Functional domains of Par-4 include two nuclear localization sequences (NLS), a leucine zipper (LZ) domain, a nuclear export sequence (NES) and selective for apoptosis in cancer cell (SAC) domain. Many studies have underlined the importance of Par-4 in preventing cancer development. The activity of Par-4 is differently regulated by localization of intracellular and extracellular Par-4. Intracellular Par-4 inhibits Akt- and NF-κB-mediated cell survival pathways and downregulates Bcl-2 expression. Extracellular Par-4 activates the extrinsic apoptotic pathway by binding to cell surface receptor GRP78, a stress response protein that is in the endoplasmic reticulum (ER). Endogenous Par-4 sensitizes cancer cells to various apoptotic stimuli, while exogenous Par-4 enhances SAC domain-dependent apoptosis in cancer cells, but not normal cells. Therefore, Par-4 is an attractive target for cancer therapy.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.4
• /
• pp.493-498
• /
• 2011

This paper presents the GraphSLAM improved by selecting the measurement with respect to their likelihoods. GraphSLAM estimates the robot's path and map by utilizing the entire history of input data. However, GraphSLAM's performance suffers a lot from severely noisy measurements. In this paper, we present GraphSLAM improved by the selective measurement method. Thus the presented GraphSLAM provides higher performance compared with the standard GraphSLAM.