Chithra Srinivasan · DR-NTU (Nanyang Technological University) · 2014
Loop closing is the problem of correctly asserting that a robot has returning to a previously visited area. This is an extremely important component of the Simultaneous Localization and Mapping (SLAM) problem (Kin Leong Ho, 2006). The many implementations of loop closure in various SLAM techniques involve the internal map and vehicle estimates to come to a conclusion. However, these implementations are prone to much error. The loop closer mechanism that is being proposed here does not involve any of the metric estimates of the SLAM system, rather it is completely independent. It uses an appearance-based SLAM technique to deal with the problem of loop closure. \n \nThe vehicle captures the appearance of the local scene with the help of a RGBD sensor like Kinect. The captured scenes are classified using a biologically inspired method which uses the shape based image property that is provided by a hierarchical feed forward model of the visual cortex. The H-MAX algorithm has been used for this purpose. The similarities of all the images through the extracted features are then encoded in a ‘similarity matrix’. These sequences as a last step are then used by a dynamic programming algorithm (Smith-Waterman algorithm) to extract the presence of a closed-loop. The technique is also demonstrated successfully with depth based images.
Cited by 0
Émilie Vlachos-Breton, Marc Drolet · Methods in molecular biology · 2022
Cited by 8
Dasong Gao, Chen Wang, Sebastian Scherer · 2022 International Conference on Robotics and Automation (ICRA) · 2022
Loop closure detection is an important building block that ensures the accuracy and robustness of simultaneous localization and mapping (SLAM) systems. Due to their generalization ability, CNN-based approaches have received increasing attention. Although they normally benefit from training on datasets that are diverse and reflective of the environments, new environments often emerge after the model is deployed. It is therefore desirable to incorporate the data newly collected during operation for incremental learning. Nevertheless, simply finetuning the model on new data is infeasible since it may cause the model's performance on previously learned data to degrade over time, which is also known as the problem of catastrophic forgetting. In this paper, we present AirLoop, a method that leverages techniques from lifelong learning to minimize forgetting when training loop closure detection models incrementally. We experimentally demonstrate the effectiveness of AirLoop on TartanAir, Nordland, and RobotCar datasets. To the best of our knowledge, AirLoop is one of the first works to achieve lifelong learning of deep loop closure detectors.
Cited by 17Open access
Wei Yuan · 2014
Sodium loop is one of the most important characteristics in computer-aided diagnosis of iris. It is related to the discretion of the cholesterol content in the body closely. This paper presents a algorithm of iris sodium loop detection under the space of HSI, Firstly, Segmenting the iris with sclera, eyelash and eyelid to complete feature selection, in order to reduce the shade of effects on the detection of sodium. Secondly, do the rectangular expansion process on iris image, to reduce the sodium loop detection range. Finally, under the HSI space, analyzing rectangular iris image's hue and saturation. Through the data fusion of H, S component to complete the sodium loop detection automatically, Experimental results show that the algorithm can achieve accurate detection of sodium loop.
Cited by 0
S Besford, G R Hedley, P M Stewart · 1981
This report describes the work performed and the results obtained under ARRB project 318 'loop detector systems in pavements', in which programmable digital techniques are used to implement an inductive loop based vehicle detector. The hardware pertaining to the tasks of driving the loop and measuring data from the loop is discussed. An algorithm which extracts vehicle related information from the measured loop data and uses this to emulate a passage vehicle detector is presented. This detector incorporates dynamic sensitivity control to minimise false detections. The concluding remarks summarise the characteristics of the digital loop detector (a).
Cited by 0
Haixia Zhai, Hao Yang, Zhanwei Hou et al. · Research Square · 2026
Cited by 0Open access
Zhe Yang, Yun Pan, Lei Deng et al. · IET Intelligent Transport Systems · 2021
Abstract Visual simultaneous localization and mapping (vSLAM), one of the most important applications in autonomous vehicles and robots to estimate the position and pose using inexpensive visual sensors, suffers from error accumulation for long‐term navigation without loop closure detection. Recently, deep neural networks (DNNs) are leveraged to achieve high accuracy for loop closure detection, however the execution time is much slower than those employing handcrafted visual features. In this paper, a parallel loop searching and verifying method for loop closure detection with both high accuracy and high speed, which combines two parallel tasks using handcrafted and DNN features, respectively, is proposed. A fast loop searching is proposed to link the bag‐of‐words features and histogram for higher accuracy, and it splits the images into multiple grids for high parallelism; meanwhile, a DNN feature extractor is utilized for further verification. A loop state control method based on a finite state machine to control these tasks is designed, wherein the loop closure detection is described as a context‐related procedure. The framework is implemented on a real machine, and the top‐2 best accuracy and fastest execution time of 80‐543 frames per second (min: 1.84ms, and max: 12.45ms) are achieved on several public benchmarks compared with some existing algorithms.
Cited by 3Open access
Sana Chouaibi, Mohamed Hadj Saïd, Antonino Imburgia et al. · Digital Access to Libraries (Université catholique de Louvain (UCL), l'Université de Namur (UNamur) and the Université Saint-Louis (USL-B)) · 2024
This article delves into the application and efficacy of planar loop sensors for detecting partial discharge (PD) phenomena. The preference for loop sensors in PD detection stems from their simple design, non-destructive testing feature, and contactless characteristics. The paper explores both experimental and finite element modeling (FEM) approaches to evaluate the performance of distinct planar square spiral sensors in PD detection. The primary objective of the study is to compare inductive loops with loop antennas by varying the geometric parameters to achieve each approach. Two single-turn square loop antennas were designed, with a circumference of 30 cm and 75 cm, and compared with four multi-turn square inductive loops with an exterior side of 31.2 mm and varying geometric parameters. The figures of merit of all sensors, such as inductance, quality factor, and resonant frequency, are extracted through measurement and simulation to identify the most influential parameter that maximizes the sensing capability toward PD.
Cited by 1Open access
V. Kettnaker, Jin Kyu Gahm · 2004
We present a system that detects people in indoor scenes by modeling the motion history of foreground blobs, rather than their shape or appearance. The system tracks all foreground blobs over time with a multi-hypothesis tracker, and considers a blob to be a person if it exhibited sufficient autonomous movement in the course of its tracking history. This way, people can be correctly classified even if they are seen in a wide range of body poses, if they remain still for a long time, or if they change appearance by taking off a coat. Evaluation on over 1h of video demonstrated good performance for both heuristic and decision tree based classification.
Cited by 1
Jacob Colton Henry · Digital WPI · 2018
Symbolic analysis of a program involves the evaluation of assembly instructions with arguments that aren’t specific constants, but instead symbolic values defined by Constraints. As instructions are executed symbolically, these constraints areupdated to reflect restrictions on what values a variable can hold, and how different variables relate to one another. On some level, the constraints upon the state of the program are a reflection of its behavior, capturing the relation between inputs and outputs as a complex mathematical expression. By using a modification of existing Frequent Pattern Mining techniques, it is possible to find repeating structures in these constraints, and from them discover characteristics of the original program’s behavior. In this report we present...
Cited by 0
Sebastian Scherer, Alina Kloss, Andreas Zell · 2013
We investigate the question whether loop closure detection using depth images is feasible using currently available depth features. For this reason, we collected a benchmark dataset consisting of a total number of 15 logfiles with several loops in various environments, implemented a modular and easily extensible loop closure detector and used this to evaluate the adequacy of state-of-the art depth features on our benchmark dataset. To allow for a fair comparison, we determined the best values for the sometimes large number of user-chosen parameters using a large-scale grid search. Since our benchmark dataset contains both depth and RGB images, we can compare the performance relying on depth features with the performance achieved when using intensity image features.
Cited by 6
Marco Baioletti, Stefano Marcugini, Alfredo Milani et al. · PORTO Publications Open Repository TOrino (Politecnico di Torino) · 1991
Cited by 0
Frieder Stolzenburg · 1999
Cited by 0
Masahiro Kitagawa, T. Kan, Takayuki Katou et al. · 1997
Cited by 0
Ying Wang, Zezhou Sun, Chengzhong Xu et al. · 2020
In this paper, a global descriptor for a LiDAR point cloud, called LiDAR Iris, is proposed for fast and accurate loop-closure detection. A binary signature image can be obtained for each point cloud after several LoG-Gabor filtering and thresholding operations on the LiDAR-Iris image representation. Given two point clouds, their similarities can be calculated as the Hamming distance of two corresponding binary signature images extracted from the two point clouds, respectively. Our LiDAR-Iris method can achieve a pose-invariant loop-closure detection at a descriptor level with the Fourier transform of the LiDAR-Iris representation if assuming a 3D (x,y,yaw) pose space, although our method can generally be applied to a 6D pose space by re-aligning point clouds with an additional IMU sensor. Experimental results on five road-scene sequences demonstrate its excellent performance in loop-closure detection.
Cited by 236
Adrien Angeli, Stéphane Doncieux, Jean-Arcady Meyer et al. · HAL (Le Centre pour la Communication Scientifique Directe) · 2008
In robotic applications of visual simultaneous localization and mapping, loop-closure detection and global localization are two issues that require the capacity to recognize a previously visited place from current camera measurements. We present an online method that makes it possible to detect when an image comes from an already perceived scene using local shape information. Our approach extends the bag of visual words method used in image recognition to incremental conditions and relies on Bayesian filtering to estimate loop-closure probability. We demonstrate the efficiency of our solution by real-time loop-closure detection under strong perceptual aliasing conditions in an indoor image sequence taken with a handheld camera.
Cited by 105Open access
Christophe Aubrun, Michel Robert, T. Cecchin · Control Engineering Practice · 1995
Cited by 7
Sabitha K S, Aashik Harishwar M L A, Elavenil B et al. · Zenodo (CERN European Organization for Nuclear Research) · 2025
The increasing complexity of cyber threats has made automated systems essential for detecting and preventing malicious activities. However, fully automated models often struggle to handle ambiguous or novel attacks that require contextual understanding. This project, Human-in-the-Loop Threat Detection and Prevention System, integrates machine learning–based detection with human expertise to create a hybrid security framework. The system continuously monitors network logs to identify anomalies such as brute-force attacks, phishing attempts, and port scans. Detected threats are automatically categorized—known attacks are handled through automated prevention mechanisms, while uncertain cases are forwarded to a web-based human interface for expert review and decision-making. Human feedback is recorded and used to retrain the model, improving its accuracy and adaptability over time. This collaborative approach bridges the gap between automation and human judgment, enabling faster, more accurate, and adaptive cyber defense against evolving threats.
Cited by 0Open access
L Chen, A D May · 1987
Cited by 0
Nurcan Durak, Olfa Nasraoui, J. T. Schmelz · Pattern Recognition · 2009
Cited by 15