Article
Mathematical & Computational Biology
Hristofor Lukanov, Peter Koenig, Gordon Pipa
Summary: The study proposes an end-to-end neural model for foveal-peripheral vision inspired by retino-cortical mapping in primates and humans, achieving high resolution for a small portion of the scene while maintaining a large field of view in low resolution. The model performs well in image and video classification tasks, reducing computational effort and memory usage significantly.
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
(2021)
Review
Computer Science, Information Systems
Wangli Hao, Ian Max Andolina, Wei Wang, Zhaoxiang Zhang
Summary: Visual information is crucial for animals' evolutionary success and has also achieved great success in computing, inspired by biological mechanisms such as deep neural networks. By surveying current work, new avenues for rethinking visual computing and designing novel neural network architectures are hoped to be offered.
FRONTIERS OF COMPUTER SCIENCE
(2021)
Article
Automation & Control Systems
Dominic Buchstaller, Jing Liu, Mark French
Summary: This note presents an alternative, simple, and self-contained proof of the deterministic properties of the Kalman filter in the discrete case, where the residuals computed by the filter are identical to the least squares disturbances. It also provides results for variations with zero and non-zero initial conditions, along with a numerical example to illustrate the deterministic properties.
INTERNATIONAL JOURNAL OF CONTROL
(2021)
Article
Computer Science, Artificial Intelligence
Junfan Wang, Yi Chen, Zhekang Dong, Mingyu Gao, Huipin Lin, Qiheng Miao
Summary: Monocular depth estimation enables machines to perceive the real world, but the prediction performance of deep learning networks can be affected by network depth and convolution operations. This paper explores the interpretability relationship between the biological visual system and monocular depth estimation networks and proposes a self-attention-based network called SABV-Depth that improves prediction accuracy. Experimental results on KITTI and NYU Depth V2 datasets demonstrate that the proposed method outperforms existing approaches in terms of prediction accuracy, object information, detail information, and edge processing.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Xiaohong Yan, Guangxin Wang, Guangyuan Wang, Yafei Wang, Xianping Fu
Summary: This paper proposes a novel method to address the low visibility problem in underwater images by imitating the color constancy mechanism in biological vision and introducing a two-pathway dehazing method. Experimental results show that the proposed method achieves better visual quality.
SIGNAL PROCESSING-IMAGE COMMUNICATION
(2022)
Article
Remote Sensing
Lin Chen, Franz Rottensteiner, Christian Heipke
Summary: In feature based image matching, distinctive features in images are detected and represented by feature descriptors. Matching is then carried out by assessing the similarity of the descriptors of potentially conjugate points. The evolution from hand-crafted feature descriptors, e.g. SIFT, to machine learning and deep learning based descriptors is discussed in detail, along with the advantages and challenges of different approaches.
GEO-SPATIAL INFORMATION SCIENCE
(2021)
Article
Chemistry, Multidisciplinary
Bokyu Kwon, Sang-il Kim
Summary: This paper proposes a recursive form of an optimal finite impulse response filter for discrete time-varying state-space models. The proposed filter is derived by employing finite horizon Kalman filtering with optimally estimated initial conditions. The optimality and unbiasedness of the filter are proved by comparison with the conventional optimal finite impulse response filter in batch form. Furthermore, an adaptive FIR filter is proposed by applying the adaptive estimation scheme to the proposed recursive optimal FIR filter.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Nikolaos Zioulis, Georgios Albanis, Petros Drakoulis, Federico Alvarez, Dimitrios Zarpalas, Petros Daras
Summary: This study introduces a biologically inspired long-range skip connection that improves the performance of UNet models in segmentation tasks, but may introduce texture transfer artifacts in dense regression tasks. The proposed HybridSkip connections show improved performance in balancing edge preservation and minimizing texture transfer artifacts.
Article
Engineering, Mechanical
Fei Yan, Wenjing Liu, Fangyan Dong, Kaoru Hirota
Summary: This study proposes a quantum-inspired online spiking neural network (QiSNN) which combines a quantum particle swarm optimization algorithm and a Kalman filtering technique to smooth and denoise the original time-series data. A novel adaptive threshold selection method is developed to determine the similarity between neurons in a repository. The experimental results demonstrate that the proposed QiSNN outperforms baseline models in predicting air quality indicators (ozone and PM10 concentrations).
NONLINEAR DYNAMICS
(2023)
Article
Mathematics, Applied
Xiaoyuan Zheng, Yu Kang, Hongyi Li, Jitao Li
Summary: In this paper, the problem of distributed filtering for a class of discrete-time nonlinear complex networks subject to the network bandwidth limitation is investigated. A solution based on the round-robin protocol and multiple description coding scheme is proposed, and the sufficient conditions to obtain the upper bound of estimation error covariance are derived.
APPLIED MATHEMATICS AND COMPUTATION
(2023)
Article
Chemistry, Analytical
Muhammad Uzair, Russell S. A. Brinkworth, Anthony Finn
Summary: Thermal infrared imaging is effective for long-range detection of small moving objects, but faces challenges such as sensor noise and cluttered backgrounds. The four-stage biologically inspired vision (BIV) model of the flying insect visual system shows excellent capability in overcoming these challenges simultaneously, outperforming existing traditional detection methods.
Article
Engineering, Mechanical
Reizinger Patrik, Vajda Ferenc
Summary: The paper proposes a Kalman filter architecture to reduce the computational cost of attitude estimation in CubeSats. The method decomposes attitude dynamics and kinematics, leading to a linear attitude quaternion and a nonlinear angular velocity filter. The virtual sensor paradigm is introduced to transform the nonlinear observation model into a linear one, without relying on approximations. Numerical experiments demonstrate superior error dynamics and robustness compared to a nonlinear quaternionic filter, and performance analysis is conducted on star tracker measurement frequency and sensitivity to the angle between Sun and Earth magnetic field measurements.
JOURNAL OF THE BRAZILIAN SOCIETY OF MECHANICAL SCIENCES AND ENGINEERING
(2023)
Article
Computer Science, Cybernetics
Phillip S. M. Skelton, Anthony Finn, Russell S. A. Brinkworth
Summary: This paper investigates the visual system of insects and their ability to navigate and avoid obstacles in complex environments. By using a novel algorithm based on nonlinear spatio-temporal feedforward filtering, the contrast dependence of time to impact estimation is overcome. This research provides a foundation for future visually guided systems.
BIOLOGICAL CYBERNETICS
(2022)
Article
Chemistry, Analytical
Ming Li, Yingjie Zhang, Zuolei Hu, Ying Zhang, Jing Zhang
Summary: The paper discusses a strategy to improve SOC estimation accuracy when battery charge and discharge conditions change, by designing an adaptive forgetting factor regression least-squares-extended Kalman filter method, and points out the advantages of this method over traditional RLS algorithm in terms of accuracy and robustness.
Article
Physics, Multidisciplinary
Meifeng Xie, Ping Zhang, Kundong Wang, Huaming Lei
Summary: Shipborne dynamic weighing is crucial for the integrated exploitation of marine fishery resources and oceanographic research. This study developed a mathematical model to analyze the impact of ship's attitude on weighing results and used compensation factors and Kalman filtering to mitigate the influence of ship oscillations on weight measurements. A shipborne dynamic weighing system was also developed and evaluated, meeting the weight measurement requirements of marine dynamic weighing and sorting.
Article
Robotics
Katerina Kalou, Giulia Sedda, Agostino Gibaldi, Silvio P. Sabatini
Summary: Interpreting three-dimensional shapes using binocular disparity and gaze information is crucial for humans and primates when exploring the surrounding environment. This research contributes to understanding the mechanisms and neural substrates of the visual system.
FRONTIERS IN ROBOTICS AND AI
(2022)
Article
Computer Science, Interdisciplinary Applications
Chiara Bassano, Manuela Chessa, Fabio Solari
Summary: This study focuses on designing and implementing a paradigm to evaluate visual working memory (VWM) in immersive visualization and developing a novel image-based computational model to mimic human VWM behavior. The study found that VWM has a capacity limit of around 7±2 items, and visual angle and observation time have an impact on VWM performance.
Proceedings Paper
Remote Sensing
Razeen Hussain, Marianna Pizzo, Giorgio Ballestin, Manuela Chessa, Fabio Solari
Summary: 3D reconstruction is a topic of interest in multiple fields, but it can be time-consuming and challenging to obtain accurate models. This study aims to provide guidance to novice users by analyzing the performance of various 3D reconstruction frameworks. The evaluation is based on synthetic data representing objects of different shapes and sizes, with metrics including mean errors and reconstruction time. The results show that Colmap performed the best in reconstruction accuracy.
2022 IEEE 5TH INTERNATIONAL CONFERENCE ON IMAGE PROCESSING APPLICATIONS AND SYSTEMS, IPAS
(2022)
Proceedings Paper
Robotics
Emanuele Riccardo Rosi, Maximilian Stolzle, Fabio Solari, Cosimo Della Santina
Summary: The nature of continuum soft robots requires novel perception solutions. This paper proposes the use of monocular cameras along with SLAM algorithm and nonlinear optimization to provide shape sensing for soft robots. The method is proven effective through simulation and experiments.
2022 IEEE 5TH INTERNATIONAL CONFERENCE ON SOFT ROBOTICS (ROBOSOFT)
(2022)
Meeting Abstract
Ophthalmology
Manuela Chessa, Lorenzo Gerini, Fabio Solari
Meeting Abstract
Ophthalmology
Manuela Chessa
Proceedings Paper
Computer Science, Cybernetics
Lorenzo Gerini, Fabio Solari, Manuela Chessa
Summary: Mixed Reality (MR) combines the visualization potential of Virtual Reality (VR) with the physical properties of the real world, allowing for a more natural interaction in virtual environments. The study shows that the cost of movements in MR is closer to the real world compared to VR, indicating that MR systems can achieve more realistic and efficient human movements.
2022 IEEE INTERNATIONAL SYMPOSIUM ON MIXED AND AUGMENTED REALITY ADJUNCT (ISMAR-ADJUNCT 2022)
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Irene Capasso, Chiara Bassano, Fabrizio Bracco, Fabio Solari, Eros Viola, Manuela Chessa
Summary: Ensuring employees receive good training in safety, risk assessment, fire prevention, and management of dangerous situations is a priority in various sectors. Virtual Reality (VR) technologies are being explored as a substitute or complement to traditional teaching methods in training courses. This study presents a multiplayer immersive VR application for safety and fire prevention training, developed in collaboration with experts using a user-centered design approach.
EXTENDED REALITY, XR SALENTO 2022, PT I
(2022)
Article
Computer Science, Information Systems
Chiara Bassano, Manuela Chessa, Fabio Solari
Summary: Exergames and serious games have gained popularity in recent years and are now being used for cognitive assessment and training. By integrating Artificial Intelligence and Web based solutions, these games can provide numerous benefits and the research in this field is still at an early stage.
Proceedings Paper
Computer Science, Cybernetics
Eros Viola, Fabio Solari, Manuela Chessa
Summary: This pilot study analyzes how users manipulate small virtual objects using different technologies, including HTC Vive controllers, Leap Motion, and Manus Prime haptic gloves. The study aims to quantitatively assess the effectiveness of these devices in pick-and-place and simple manipulation tasks. The results show that the gloves have tracking issues with the thumb, while the controllers are a good compromise and the Leap Motion's vision-based solution is appreciated by users.
PROCEEDINGS OF THE 17TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS (HUCAPP), VOL 2
(2022)
Article
Computer Science, Artificial Intelligence
Chenglong Li, Xiaobin Yang, Guohao Wang, Aihua Zheng, Chang Tan, Jin Tang
Summary: License plate recognition is crucial in various practical applications, however, recognizing license plates of large vehicles is challenging due to low resolution, contamination, low illumination, and occlusion. To address this problem, a novel data generation framework based on the Disentangled Generation Network is proposed to ensure the generation diversity and integrity for robust enlarged license plate recognition.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Sara Casao, Alvaro Serra-Gomez, Ana C. Murillo, Wendelin Bohmer, Javier Alonso-Mora, Eduardo Montijano
Summary: This paper presents a hybrid camera system that combines static and mobile cameras, exploiting the cooperation between tracking and control modules to achieve high-level scene understanding. The static camera network provides global awareness, while the mobile cameras enhance the information about the people on the scene.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Anh-Dzung Doan, Bach Long Nguyen, Surabhi Gupta, Ian Reid, Markus Wagner, Tat-Jun Chin
Summary: To ensure reliable object detection in autonomous systems, the detector needs to adapt to changes in appearance caused by environmental factors. We propose a selective adaptation approach using domain gap as a criterion to improve the efficiency of the detector's operation.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Tianhong Dai, Wei Li, Xilei Cao, Jianzhuang Liu, Xu Jia, Ales Leonardis, Youliang Yan, Shanxin Yuan
Summary: This study proposes a novel frequency-guided deep neural network (FHDRNet) for high dynamic range (HDR) imaging from multiple low dynamic range (LDR) images, aiming to address ghosting artifacts. By conducting HDR fusion in the frequency domain, the network utilizes low-frequency signals to remove specific ghosting artifacts and high-frequency signals to preserve details. Extensive experiments demonstrate that this approach achieves state-of-the-art performance.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Guobin Li, Reyer Zwiggelaar
Summary: Breast cancer is the most commonly diagnosed female malignancy worldwide. Recent developments in deep convolutional neural networks have shown promising performance for breast cancer detection and classification. However, biased features can be learned due to variations in appearance and small datasets. To address this issue, a densely connected convolutional network (DenseNet) was trained using texture features representing different physical morphological representations as inputs. The use of connectivity estimation and nearest neighbors improved the network's unbiased prediction. The approach achieved higher diagnostic accuracy and provided visual explanations for model predictions.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Yuezun Li, Cong Zhang, Honggang Qi, Siwei Lyu
Summary: Deep Neural Networks (DNNs) are vulnerable to adversarial perturbations, limiting their applicability in safe-critical scenarios. To address this, a new method called AdaNI is proposed to increase feature randomness through adaptive noise injection, improving adversarial robustness. Extensive experiments demonstrate the efficacy of AdaNI against various white-box and black-box attacks, as well as its applicability in DeepFake detection.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Chengyin Hu, Weiwen Shi, Ling Tian, Wen Li
Summary: In this study, we introduce a pioneering black-box light-based physical attack called Adversarial Neon Beam (AdvNB). Our method excels in attack modeling, efficient attack simulation, and robust optimization, striking a balance between robustness and efficiency. Through rigorous evaluation, we achieve impressive attack success rates in both digital and real-world scenarios. AdvNB demonstrates its stealthiness through comparisons with baseline samples and consistently achieves high success rates when targeting advanced DNN models.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Hang Wang, Zhenyu Ding, Cheng Cheng, Yuhai Li, Hongbin Sun
Summary: Learning-based super resolution has made remarkable progress in improving image quality, but the performance decreases when the degradation kernel changes. Blind SR networks can estimate the degradation kernel and adapt well in realistic scenarios, improving performance and runtime. This paper proposes a design that imposes constraints for the kernel estimation network in both the image domain and kernel domain, resulting in high-quality images and efficient runtime.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Yuantao Chen, Runlong Xia, Kai Yang, Ke Zou
Summary: This paper proposes an improved image inpainting network using a multi-scale feature module and improved attention module. The network addresses issues in deep learning-based image inpainting algorithms, such as information loss in deep level features and the neglect of semantic features. The proposed network generates better inpainting results by reducing information loss and enhancing the ability to restore texture and semantic features.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
Article
Computer Science, Artificial Intelligence
Yi-Tung Chan
Summary: This study proposes a novel maritime background subtraction method based on ensemble learning theory to address the challenges posed by dynamic marine environments and noise, improving the detection accuracy and enhancing maritime transportation security for autonomous ships in open waters.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2024)
