Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks314 from public data
- A Comprehensive Overview of Large Language Models2,059
This review article is intended to provide a quick, comprehensive reference for the researchers and practitioners to draw insights from extensive, informative summaries of the existing works to advance the LLM research.
- Real Image Denoising With Feature Attention719
A novel single-stage blind real image denoising network (RIDNet) is proposed by employing a modular architecture that uses residual on the residual structure to ease the flow of low-frequency information and apply feature attention to exploit the channel dependencies.
- Simultaneously Localize, Segment and Rank the Camouflaged Objects599
It is argued that explicitly modeling the conspicuousness of camouflaged objects against their particular backgrounds can not only lead to a better understanding about camouflage and evolution of animals, but also provide guidance to design more sophisticated camouflage techniques.
- A Deep Journey into Super-resolution503
A taxonomy for deep learning–based super-resolution networks is introduced that groups existing methods into nine categories including linear, residual, multi-branch, recursive, progressive, attention-based, and adversarial designs.
- UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders398
Inspired by the saliency data labeling process, a probabilistic RGB-D saliency detection network via conditional variational autoencoders to model human annotation uncertainty and generate multiple saliency maps for each input image by sampling in the latent space is proposed.
- 371
- First-in-Human Trial of a Novel Suprachoroidal Retinal Prosthesis330
The first-in-human Phase 1 trial to investigate the use of retinal implants in the suprachoroidal space in three human subjects with end-stage retinitis pigmentosa holds promise for the field of vision restoration.
- 300
- Densely Residual Laplacian Super-Resolution288
This work presents a compact and accurate super-resolution algorithm, namely, densely residual laplacian network (DRLN), which employs cascading residual on the residual structure to allow the flow of low-frequency information to focus on learning high and mid-level features.
- Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion275
This work comprehensively interpret the distinctness of the points from multiple resolutions and represent the feature map following an adaptive fusion method at point-level for accurate semantic segmentation for large-scale point cloud data collected in reality.
- Reducing the Sim-to-Real Gap for Event Cameras262
This work presents a new High Quality Frames (HQF) dataset, containing events and ground truth frames from a DAVIS240C that are well-exposed and minimally motion-blurred, and presents strategies for improving training data for event based CNNs that result in 20-40% boost in performance of existing state-of-the-art (SOTA) video reconstruction networks retrained with this method.
- 260
- Continuous-Time Intensity Estimation Using Event Cameras254
A computationally efficient, asynchronous filter that continuously fuses image frames and events into a single high-temporal-resolution, high-dynamic-range image state is proposed that outperforms existing state-of-the-art methods.
- An update on retinal prostheses221
Over 40 individual research groups who have built devices, developed stimulation strategies, or investigated the basic physiology underpinning retinal prostheses are encompassed, along with future challenges that face the field.
- 216
- 186
- 181
- Uncertainty Inspired RGB-D Saliency Detection168
Qualitative and quantitative results on six challengingRGB-D benchmark datasets show the first stochastic framework to employ uncertainty for RGB-D saliency detection by learning from the data labeling process superior performance in learning the distribution of saliency maps.
- RGB-D Saliency Detection via Cascaded Mutual Information Minimization131
This paper introduces a novel multistage cascaded learning framework via mutual information minimization to explicitly model the multi-modal information between RGB image and depth data.
- Improved Visual-Semantic Alignment for Zero-Shot Object Detection125
An end-to-end deep learning framework underpinned by a novel loss function that handles class-imbalance and seeks to properly align the visual and semantic cues for improved zero-shot learning is proposed.
- Dense-Resolution Network for Point Cloud Classification and Segmentation122
A novel network named Dense-Resolution Network (DRNet) is proposed for point cloud analysis designed to learn local point features from the point cloud in different resolutions and presents a novel grouping method for local neighborhood searching and an error-minimizing module for capturing local features.
- Toward Deeper Understanding of Camouflaged Object Detection115
This paper argues that the binary segmentation setting fails to fully understand the concept of camouflage, and presents the first triple-task learning framework to simultaneously localize, segment, and rank camouflaged objects, indicating the conspicuousness level of camouflage.
- CED: Color Event Camera Dataset111
This work presents and releases the first Color Event Camera Dataset (CED), containing 50 minutes of footage with both color frames and events, and presents an extension of the event camera simulator ESIM that enables simulation of color events.
- 108
- Correlating driver gaze with the road scene for driver assistance systems107
This paper presents a complete system that reads speed signs in real-time, compares the driver's gaze, and provides immediate feedback if it appears the sign has been missed by the driver.
- Towards Open-Set Object Detection and Discovery104
This work proposes a two-stage method that first uses an open-set object detector to predict both known and unknown objects, then studies the representation of predicted objects in an unsupervised manner and discover new categories from the set of unknown objects.
- PU-Transformer: Point Cloud Upsampling Transformer97
To activate the transformer's strong capability in representing features, a new variant of a multi-head self-attention structure is developed to enhance both point-wise and channel-wise relations of the feature map.
- Weakly Supervised Video Salient Object Detection89
An "Appearance-motion fusion module" and bidirectional ConvLSTM based framework are proposed to achieve effective multi-modal learning and long-term temporal context modeling based on the first weakly supervised video salient object detection model based on relabeled “fixation guided scribble annotations”.
- 87
- 86
- 84
- 84
- Learning RGB-D Salient Object Detection Using Background Enclosure, Depth Contrast, and Top-Down Features80
This work proposes a novel deep CNN architecture for RGB-D salient object detection that utilizes both top-down and bottom-up cues and presents novel depth features that capture the ideas of background enclosure, depth contrast and histogram distance in a manner that is suitable for a learned approach.
- Learning Audio-Visual Source Localization via False Negative Aware Contrastive Learning76
This work proposes a new learning strategy named False Negative Aware Contrastive (FNAC), which utilizes the intra-modal similarities to identify potentially similar samples and construct corresponding adjacency matrices to guide contrastive learning.
- 68
- 66
- Omni-drive Robot Motion on Curved Paths: The Fastest Path between Two Points Is Not a Straight-Line66
It is shown that for three-wheeled omni-drive transport systems and certain ranges of trajectories and starting conditions, a curved path can be traversed faster than a straight-line path, and this result is confirmed experimentally.
- Learning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency Detection64
A noise-aware encoder-decoder framework to disentangle a clean saliency predictor from noisy training examples, where the noisy labels are generated by unsupervised handcrafted feature-based methods is proposed.
- Asynchronous Spatial Image Convolutions for Event Cameras64
A method to compute the convolution of a linear spatial kernel with the output of an event camera, providing an implementation of a Harris corner-response “state” that can be used in real time for feature detection and tracking on robotic systems.
- Impacts of impaired face perception on social interactions and quality of life in age-related macular degeneration: A qualitative study and new community resources64
Poor face perception in AMD is an important domain contributing to impaired social interactions and quality of life, and this domain should be directly assessed in quantitativequality of life measures, and in resources designed to improve community understanding.
- Computing Spatial Image Convolutions for Event Cameras.64
A method to compute the convolution of a linear spatial kernel with the output of an event camera, providing an implementation of a Harris corner-response “state” that can be used in real time for feature detection and tracking on robotic systems.
- A Second-Generation (44-Channel) Suprachoroidal Retinal Prosthesis: Interim Clinical Trial Results61
These interim study data demonstrate that the suprachoroidal prosthesis is safe and provides significant improvements in functional vision, activities of daily living, and observer-rated quality of life.
- Vicinity Vision Transformer58
Vicinity Attention, which is a type of linear attention that integrates 2D locality that achieves state-of-the-art image classification accuracy with 50% fewer parameters than previous approaches, is proposed.
- PnP-3D: A Plug-and-Play for 3D Point Clouds51
A plug-and-play module, PnP-3D, aiming to refine the fundamental point cloud feature representations by involving more local context and global bilinear response from explicit 3D space and implicit feature space is proposed.
- Knowledge-Based Autonomous Dynamic Colour Calibration50
Colour labeling is critical to the real-time performance of colour-based vision systems and is used for low-level vision by most RoboCup 2002 physically based teams but manual calibration is both time consuming and error prone.
- 49
- P2C: Self-Supervised Point Cloud Completion from Single Partial Clouds47
This work presents Partial2Complete (P2C), the first self-supervised framework that completes point cloud objects using training samples consisting of only a single incomplete point cloud per object, and proposes Region-Aware Chamfer Distance to regularize shape mismatch without limiting completion capability.
- Spectral-GANs for High-Resolution 3D Point-cloud Generation47
This paper develops a principled approach to synthesize 3D point-clouds using a spectral-domain Generative Adversarial Network (GAN) and proposes a fully differentiable block to transform from the spectral to the spatial domain and back, thereby allowing us to integrate knowledge from well-established spatial models.
- Mobility and low contrast trip hazard avoidance using augmented depth47
This is the first demonstration that an augmentation of the scene ensuring key objects are visible may provide better outcomes for prosthetic vision, and indicates that augmented depth may enable safe mobility in the presence of low-contrast obstacles with current and near-term implants.
- 46
- 45
- Vision function testing for a suprachoroidal retinal prosthesis: effects of image filtering45
The first comprehensive results of vision function testing for a suprachoroidal retinal prosthetic device utilizing of 20 stimulating electrodes are presented and vision processing with a Nyquist bandlimited image filter has shown an advantage for a light localization task.
- 44
- 43
- Visual task performance in the blind with the BrainPort V100 Vision Aid42
In this work, the functional studies performed to date with the BrainPort are reviewed and the critical factors that determine device performance are highlighted, including the technology behind the brainPort, the impediments to assessing device performance, and the impact of device training and rehabilitation.
- Deep Texture and Structure Aware Filtering Network for Image Smoothing41
This paper generates a large dataset by blending natural textures with clean structure-only images, and uses this to build a texture prediction network (TPN) that predicts the location and magnitude of textures, and combines the TPN with a semantic structure predictions network (SPN) to form the final texture and structure aware filtering network (TSAFN).
- 41
- 40
- An Alternative to WSSS? An Empirical Study of the Segment Anything Model (SAM) on Weakly-Supervised Semantic Segmentation Problems38
This report adapts SAM as the pseudo-label generation pipeline given only the image-level class labels in Weakly-Supervised Semantic Segmentation (WSSS), and observes impressive results in most cases.
- Unsupervised Primitive Discovery for Improved 3D Generative Modeling38
A novel factorized generative model for 3D shape generation that sequentially transitions from coarse to fine scale shape generation and an unsupervised primitive discovery algorithm based on a higher-order conditional random field model are introduced.
- 38
- 38
- Identification of Characters and Localization of Images Using Direct Multiple-Electrode Stimulation With a Suprachoroidal Retinal Prosthesis37
The ability to identify characters and localize percepts demonstrates the capacity of the suprachoroidal device to provide meaningful information to blind patients and highlights the importance of using spatial cues from phosphenes, which becomes more difficult at low contrast.
- The regular polygon detector37
A robust regular polygon detector that shows robustness to noise, the ability to find and differentiate different shape types, and to perform real-time sign detection for driver assistance is described.
- 36
- 36
- Face identity recognition in simulated prosthetic vision is poorer than previously reported and can be improved by caricaturing35
Results show caricaturing improved identity recognition in memory and/or perception (degree by which two faces look dissimilar) down to a resolution of 32 × 32 with 30% phosphene dropout, imply caricatured may offer benefits for patients at resolutions realistic for some current‐generation or in‐development implants.
- 33
- 32
- Learning Saliency From Single Noisy Labelling: A Robust Model Fitting Perspective31
This paper calls on the theory of robust model fitting and formulate deep saliency prediction from a single noisy labelling as robust network learning and exploit model consistency across iterations to identify inliers and outliers.
- The role of visual deprivation and experience on the performance of sensory substitution devices31
Establishing how experience and different etiologies of vision loss affect the performance of these devices may help to improve existing rehabilitation strategies, formulate effective selection criteria and develop prognostic measures is discussed.
- Transmission-Guided Bayesian Generative Model for Smoke Segmentation30
A transmission-guided local coherence loss is designed to guide the network to learn pair-wise relationships based on pixel distance and the transmission feature to effectively model the two types of uncertainty in smoke segmentation.
- 30
- Attention-Based Real Image Restoration30
A novel single-stage blind real image restoration network is proposed by employing a modular architecture that uses a residual on the residual structure to ease low-frequency information flow and apply feature attention to exploit the channel dependencies.
- Semantic labeling for prosthetic vision30
This work introduces semantic labeling as a technique to improve navigation outcomes for prosthetic vision users, and produces a novel egocentric vision dataset to demonstrate how semantic labeling can be applied to this problem.
- 30
- Adversarial Training of Variational Auto-Encoders for High Fidelity Image Generation29
This model simultaneously learns to match the data, reconstruction loss and the latent distributions of real and fake images to improve the quality of generated samples and uses an error feedback approach to maintain the equilibrium between competing networks in the model.
- 28
- 28
- 27
- 27
- 26
- 26
- 26
- Perspective Rectangle Detection26
The detector has been designed for robot visual mapping, and the application of this detector to real-time stereo matching and reconstruction over a corridor sequence for this goal is presented.
- 25
- 24
- Weakly-Supervised Contrastive Learning for Unsupervised Object Discovery23
A semantic-guided self-supervised learning model to extract high-level semantic features from images is designed by fine-tuning the feature encoder of a self-supervised model, namely DINO, via WCL, and Principal Component Analysis (PCA) is introduced to localize object regions.
- A Second-Generation (44-Channel) Suprachoroidal Retinal Prosthesis: A Single-Arm Clinical Trial of Feasibility23
The completed clinical study demonstrates that the suprachoroidal prosthesis raises no safety concerns and provides improvements in functional vision, activities of daily living, and observer-rated quality of life.
- Determining the Contribution of Retinotopic Discrimination to Localization Performance With a Suprachoroidal Retinal Prosthesis23
Investigation of the contribution of retinotopic electrode discrimination to visual performance for three recipients of a 24-channel suprachoroidal retinal implant found successful discrimination of ret inotopic information is possible with a suprACHoroidalretinal prosthesis.
- 23
- Training Improves Vibrotactile Spatial Acuity and Intensity Discrimination on the Lower Back Using Coin Motors21
It was concluded that short, automated training sessions could provide a fast and inexpensive means to improve people's basic spatial acuity and intensity discrimination skills with coin motors.
- A new theoretical approach to improving face recognition in disorders of central vision: Face caricaturing21
This work reports caricature advantages for all blur levels, for face viewpoints from front view to semiprofile, and in tasks involving perceiving differences in facial identity between pairs of people, remembering previously learned faces, and rejecting new faces as unknown.
- Hippocampal shape analysis for Alzheimer's disease using an efficient hypothesis test and regularized discriminative deformation21
A framework to perform statistical shape analysis for segmented hippocampi is presented, including an efficient permutation test for detecting subtle class differences, and a regularized discriminative direction method for visualizing the shape discrepancy.
- 20
- Caricaturing as a general method to improve poor face recognition: Evidence from low-resolution images, other-race faces, and older adults.20
It is shown that caricaturing faces-that is, exaggerating their appearance away from an average face-can provide a useful applied method for improving face recognition across all circumstances including low-resolution images, other-race faces, and in older adult observers.
- MAP ZDF segmentation and tracking using active stereo vision: Hand tracking case study20
A maximum a posterior probability zero disparity filter (MAP ZDF) ensures coordinated stereo fixation upon an arbitrarily moving, rotating, re-configuring hand, performing marker-less pixel-wise segmentation of the hand.
- Direction Control for an Active Docking Behaviour Based on the Rotational Component of Log-Polar Optic Flow20
An algorithm and implementation for a special case of the docking problem for ground-based robots that requires the robot to dock with a fixated environment point where only visual information is available, and is suited to real-time implementation.
- Frontiers in Intelligent Colonoscopy19
This study investigates the frontiers of intelligent colonoscopy techniques and their prospective implications for multimodal medical applications by assessing the current data-centric and model-centric landscapes through four tasks for colonoscopic scene perception, including classification, detection, segmentation, and vision-language understanding.
- 19
- Inferring the Class Conditional Response Map for Weakly Supervised Semantic Segmentation19
The class conditional inference-time approach is presented to separately and iteratively reveal the classification network’s hidden object activation to generate more complete response maps and introduce a novel way to exploit saliency maps during segmentation training and refine the foreground object masks without suppressing background objects.
- Any-Shot Object Detection19
A unified any-shot detection model that can concurrently learn to detect both zero-shot and few-shot object classes, and use class semantics as prototypes for object detection, a formulation that naturally minimizes knowledge forgetting and mitigates the class-imbalance in the label space.
- Accuracy vs. complexity: A trade-off in visual question answering models19
This work systematically study the trade-off between the model complexity and the performance on the VQA task, and specifically focuses on the effect of "multi-modal fusion" in V QA models that is typically the most expensive step in a VZA pipeline.
- 19
- Oculomotor Responses to Dynamic Stimuli in a 44-Channel Suprachoroidal Retinal Prosthesis18
Retinal implant users can make naturalistic eye movements in response to moving stimuli, highlighting the potential for eye tracker feedback to improve perceptual localization and image stabilization in camera-based visual prostheses.
- Towards Safer Roads by Integration of Road Scene Monitoring and Vehicle Control18
A framework for interactive driver assistance systems that includes techniques for fast-speed sign detection and classification, obstacle detection and tracking applied to pedestrian detection, and lane departure warning is presented.
- 18
- All-pairs Consistency Learning for Weakly Supervised Semantic Segmentation17
A new transformer-based regularization to better localize objects for Weakly supervised semantic segmentation (WSSS) and introduces a novel class-wise localization method that leverages the gradients of the class token.
- 17
- KNOWLEDGE-BASED SHAPE-FROM-SHADING17
By applying domain specific knowledge, this approach can fully automate the derivation of the approximate shape of an object, and can yield specific advantages over existing approaches, both in terms of computation and processing results.
- 16
- Maximal Cliques Based Rigid Body Motion Segmentation with a RGB-D Camera16
This paper presents a novel method for motion segmentation without explicitly estimating motions, and shows that identifying true motions from the candidate motions can be cast as a minimum set cover problem or a max k-cover problem for data with outliers.
- Active Vision - Rectification and Depth Mapping16
The benefits of using an occupancy grid framework for the fusion and representation of range data, and find that it is especially suited for active vision, are outlined.
- 15
- Energy-Based Generative Cooperative Saliency Prediction15
A generative cooperative saliency prediction framework, where a conditional latent variable model and a conditional energy-based model are jointly trained to predict salient objects in a cooperative manner, which can produce a set of diverse and plausible saliency maps of an image.
- Energy-Based Residual Latent Transport for Unsupervised Point Cloud Completion15
This work proposes a novel framework that performs completion by transforming a partial shape encoding into a complete one using a latent transport module, and it is designed as a latent-space energy-based model (EBM) in an encoder-decoder architecture, aiming to learn a probability distribution conditioned on the partial shape encode.
- 3D Guided Weakly Supervised Semantic Segmentation15
This paper proposes a weakly supervised 2D semantic segmentation model by incorporating sparse bounding box labels with available 3D information, which is much easier to obtain with advanced sensors.
- 15
- From known to the unknown: Transferring knowledge to answer questions about novel visual and semantic concepts15
This work proposes an exemplar-based approach that transfers learning from previously `Known' concepts to answer questions about the `Unknown', and introduces an efficient matching scheme that uses a compact feature description for search and retrieval.
- 15
- Improving face identity perception in age-related macular degeneration via caricaturing15
It is concluded caricaturing has the potential to help many AMD patients recognise faces and was effective in half the eyes with moderate and severe vision loss.
- 15
- Knowledge Based Shape from Shading15
This paper demonstrates the efficacy of using knowledge in the basic operation of an early vision operator, and introduces a new paradigm for computer vision that may be applied to other early vision operators.
- Measuring and Modeling Uncertainty Degree for Monocular Depth Estimation14
This paper proposes to model the uncertainty of MDE models from the perspective of the inherent probability distributions originating from the depth probability volume and its extensions, and to assess it more fairly with more comprehensive metrics.
- Rethinking Polyp Segmentation From An Out-of-distribution Perspective14
This work rethink colorectal polyp segmentation from an out-of-distribution perspective with a simple but effective self-supervised learning approach that has excellent segmentation performance and generalises across datasets.
- 14
- Caricaturing faces to improve identity recognition in low vision simulations: How effective is current-generation automatic assignment of landmark points?14
The 68-point caricatures produced significant improvements in identity discrimination relative to veridical, and were approximately 50% as effective as the hand-assigned 147- point caricatures.
- 14
- 14
- 14
- Estimating Phosphene Locations Using Eye Movements of Suprachoroidal Retinal Prosthesis Users13
Gaze-based phosphene mapping was achievable in two of three tested retinal prosthesis subjects and their derivedphosphene maps correlated well with the retinotopic electrode layout, but this may have revealed that their phosphenes were indistinct spatially.
- Caricaturing can improve facial expression recognition in low-resolution images and age-related macular degeneration13
Testing potential benefits of caricaturing (exaggerating how the expression's shape differs from neutral) as an image enhancement procedure targeted at mid- to high-level cortical vision found to emerge when improvement was most needed.
- VISION GUIDED CIRCUMNAVIGATING AUTONOMOUS ROBOTS13
A system for vision guided autonomous circumnavigation, allowing robots to navigate around objects of arbitrary pose using a canonical viewer-centred model, which can be used in many applications where robots have to recognize and manipulate objects of unknown pose and placement.
- PointCaM: Cut-and-Mix for open-set point cloud learning12
This paper proposes a novel Point Cut-and-Mix mechanism for solving open-set point cloud learning, comprising an Unknown-Point Simulator and an Unknown-Point Estimator module, which leverages multi-level feature contexts to recognize unknown point cloud objects more effectively.
- Learning Camera Pose from Optical Colonoscopy Frames Through Deep Convolutional Neural Network (CNN)12
It is shown that transfer learning from surgical simulation to actual endoscopic based surgery is a possible approach for deep learning technologies and could be estimated with higher accuracy and speed than feature based computer vision methods such as the classical structure from motion pipeline.
- 12
- Efficient active monocular fixation using the log-polar sensor12
An important benefit of the log-polar sensor with vote-based algorithms, an emphasis of foveal pixels is demonstrated, which removes the need for searching explicitly through all detected corner positions in order to locate the most central.
- A Systematic Evaluation: Fine-Grained CNN vs. Traditional CNN Classifiers11
This work investigates the performance of landmark traditional CNN classifiers, presenting top-notch results on large-scale classification datasets and comparing them against state-of-the-art fine-grained classifiers.
- 11
- Question-Agnostic Attention for Visual Question Answering11
This paper proposes a question-agnostic attention mechanism that is complementary to the existing question-dependent attention mechanisms, and shows that incorporating complementary QAA allows state-of-the-art VQA models to perform better, and provides significant boost to simplistic V QA models, enabling them to performance on par with highly sophisticated fusion strategies.
- 11
- OfGAN: Realistic Rendition of Synthetic Colonoscopy Videos11
The Optical Flow Generative Adversarial Network (OfGAN) is developed to transform simulated colonoscopy videos into realistic ones while preserving annotation, and it is demonstrated that the performance of the OfGAN overwhelms the baseline method in relative tasks through both qualitative and quantitative evaluation.
- 11
- 11
- Insect Inspired Robots11
This paper introduces a research platform, the Insectbot, that was built as a testbed for new biologically-inspired, vision-based algorithms and systems, and presents the navigational behaviours currently implemented on this robot.
- Embodied categorisation for vision-guided mobile robots11
A philosophical and psycho-physiological basis for embodied perception is outlined, and a framework for conceptual embodiment of vision-guided robots is developed, and it is demonstrated that embodied categorisation can make algorithms more efficient and robust.
- 10
- 10
- 10
- 10
- LAM3D: Large Image-Point-Cloud Alignment Model for 3D Reconstruction from Single Image9
A novel framework, the Large Image and Point Cloud Alignment Model (LAM3D), which utilizes 3D point cloud data to enhance the fidelity of generated 3D meshes and achieves state-of-the-art high-fidelity 3D mesh reconstruction from a single image in just 6 seconds.
- 9
- Representation Learning on Unit Ball with 3D Roto-translational Equivariance9
A theoretical framework for volumetric convolution based on Zernike polynomials is developed and efficiently implement it as a differentiable and an easily pluggable layer in deep networks.
- 9
- Surface Extraction from Iso-disparity Contours9
The advantages of this approach over a range of indoor and outdoor stereo images, and show that iso-disparity analysis can provide a robust and efficient means of segmenting smooth surfaces, and obtaining planar surface models.
- 9
- Asymmetric Totally-Corrective Boosting for Real-Time Object Detection9
New boosting algorithms are developed that explicitly optimize asymmetric loss objectives in a totally corrective fashion and outperform the state-of-the-art asymmetric boosting methods.
- 9
- 8
- Learning structure-aware semantic segmentation with image-level supervision8
This paper argues that the lost structure information in CAM limits its application in downstream semantic segmentation, leading to deteriorated predictions, and introduces an auxiliary semantic boundary detection module, which penalizes the deteriorated predictions.
- Unsupervised Learning of Endoscopy Video Frames’ Correspondences from Global and Local Transformation8
This paper presents an unsupervised learning method to estimate dense image correspondences (DIC) between endoscopy frames by developing a new CNN model, called the EndoRegNet, which has three distinguishing aspects: a local DIC estimator, a polynomial image transformer which regularizes local correspondences and a visibility mask which refines image correspondence.
- 8
- 8
- 8
- The role of computer vision in prosthetic vision8
Some barriers to restoring functional human vision for current generation visual prosthetic devices that computer vision can help overcome are highlighted.
- 8
- 8
- Performance of temporal filters for optical flow estimation in mobile robot corridor centring and visual odometry8
The results show the best performing filter for use in the control loop is a recursive temporal filter, out performing the traditionally used Gaussian filter.
- 8
- Model Calibration in Dense Classification with Adaptive Label Perturbation7
Adaptive Stochastic Label Perturbation (ASLP) is proposed which learns a unique label perturbation level for each training image and improves model calibration degree by minimising the gap between the prediction accuracy and expected confidence of the target training label.
- How to Train Your Event Camera Neural Network7
This work presents a new High Quality Frames (HQF) dataset, containing events and groundtruth frames from a DAVIS240C that are well-exposed and minimally motion-blurred, and presents strategies for improving training data for event based CNNs that result in 25-40% boost in performance of existing state-of-the-art (SOTA) video reconstruction networks retrained with this method.
- 7
- 7
- 7
- A Study of Hippocampal Shape Difference Between Genders by Efficient Hypothesis Test and Discriminative Deformation7
The Fisher permutation and bootstrap tests are used to differentiate hippocampal shape between genders and a "informative direction" is found to preserve the original discriminative information.
- Model-Image Registration of Parametric Shape Models: Fitting a Shell to the Cochlea7
A Model-Image Registration framework that can be used to evaluate models of anatomical shape and show promising results on registering with the clinical data is presented.
- Regular Polygon Detection as an Interest Point Operator for SLAM7
Results from the application of the regular polygon detector as an interest point operator on a robot camera sequence in an indoor office environment show good results for non-trivial frame baselines.
- Towards Real-Time Strategic Teamwork: A RoboCup Case Study7
An architectural overview of the system involving an integration of an agent-oriented programming framework to support strategic decisions, with various low-level perception and control elements is presented.
- 7
- Learning colon centreline from optical colonoscopy, a new way to generate a map of the internal colon surface6
A novel convolutional neural network (ConvNet) algorithm is introduced to map the internal colon surface to a 2D map (visibility map), which can be used to increase the awareness of clinicians about areas they might miss.
- Psychophysical Evaluation of a Tactile Display Based on Coin Motors6
This study has tested the performance of a prototype 96-tactor vibrotactile using a subset of 64 tactors and determined the tactile spatial acuity and intensity discrimination in 14 naïve subjects.
- 6
- 6
- 6
- 6
- Totally-Corrective Multi-class Boosting6
This work proposes a column-generation based totally-corrective framework for multi-class boosting learning by looking at the Lagrange dual problems and proposes two boosting algorithms that have comparable generalization capability but converge much faster than their counterparts.
- Humanoid Vision Resembles Primate Archetype6
This work hypothesise that similarities between the underlying robotic system model and that of the primate vision system will elicit accordingly similar gaze behaviours, and presents a robotic vision system that incorporates these competencies.
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- Hippocampal Shape Classification Using Redundancy Constrained Feature Selection5
This work identifies the drawbacks of SVM-RFE, and proposes a novel class-separability-based feature selection approach to overcome them, and formulate feature selection as a constrained integer optimization and develop a new algorithm to efficiently and optimally solve this problem.
- 5
- On Real-Time Synthetic Primate Vision5
It is hypothesize that similarities between the underlying synthetic system model and that of the primate vision system elicit accordingly similar gaze behaviors and presents a synthetic vision system that incorporates these competencies.
- Topological Repair on Voxel-Based Quadrangular Meshes5
This work proposes a new approach to repair a voxel-based quadrangular mesh so that it becomes topologically equivalent to a sphere, which was successfully applied to a database of 977 hand-segmented hippocampi from a study of ageing.
- Sign Detection and Driving Competency for Older Drivers with Impaired Vision5
Whether driver assistance systems that are currently the subject of research can help in identifying deficits that lead to older adults needing to cease driving, and can technologies reduce the impact of these deficits on driver competence are examined.
- 5
- 4
- Transferable Attack for Semantic Segmentation4
This work proposes an ensemble attack for semantic segmentation to achieve more effective attacks with higher transferability, and finds two main factors to achieve transferable attack.
- 4
- Blended Convolution and Synthesis for Efficient Discrimination of 3D Shapes4
A two-pronged solution to this problem that is seamlessly integrated in a single blended convolution and synthesis layer that achieves compelling results on 3D shape recognition and retrieval is proposed.
- Geometric Feedback Network for Point Cloud Classification.4
A network designed as a feedback mechanism, a procedure allowing the modification of the output via as a response to the output, to comprehensively capture the local features of 3D point clouds is proposed.
- Learning Hough Transform with Latent Structures for Joint Object Detection and Pose Estimation4
A novel max-margin Hough transform with latent structure for joint object detection and pose estimation and adopts a multiple instance learning approach to handle the lack of correspondence among training instances with noisy bounding-box labels.
- 4
- 4
- 4
- 4
- Face Detection And Tracking In Video To Facilitate Face Recognition In A Visual Prosthesis4
This work is developing a real-time system capable of automatically detecting, tracking and zooming onto faces, and facilitating face and expression recognition with low-resolution retinal implant array.
- 4
- Regularized Discriminative Direction for Shape Difference Analysis4
A regularized discrim inative direction is developed by requiring a shape to conform to its population distribution when it deforms along the discriminative direction, and shows superior performance in detecting and localizing the difference of hippocampal shapes for sex.
- On-Board Vision Using Visual-Servoing for RoboCup F-180 League Mobile Robots4
This paper will present an alternative implementation of image-based servoing for approaching objects on the ground plane, which disposes of the pose estimation step and fully relies only on image features.
- 4
- 4
- 4
- Face and Available Chair Detection and Localization With a Second-Generation (44-Channel) Suprachoroidal Retinal Prosthesis3
FaD and ChD VP methods performed better than L2 for the purpose of localizing faces and available chairs, respectively and can improve the localization of specific object types while using the second-generation suprachoroidal retinal prosthesis.
- An Energy-Based Prior for Generative Saliency3
Experimental results show that the proposed generative saliency model with an energy-based prior can achieve not only accurate saliency predictions but also reliable uncertainty maps that are consistent with human perception.
- 3
- Efficient Gaussian Process Model on Class-Imbalanced Datasets for Generalized Zero-Shot Learning3
A Neural Network model that learns a latent feature embedding and a Gaussian Process regression model that predicts latent feature prototypes of unseen classes is proposed that can achieve state-of-the-art (SOTA) performance on imbalanced ZSL benchmark datasets.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- Cost modelling — an integrated approach to planning and cost control3
The mechanics of project control decisions are explored and the superiority of systems which concentrate on forecasting is demonstrated to show why forecasting based on computer-aided cost modelling are effective.
- NumGrad-Pull: Numerical Gradient Guided Tri-Plane Representation for Surface Reconstruction From Point Clouds2
NumGrad-Pull is introduced, leveraging the representation capability of tri-plane structures to accelerate the learning of signed distance functions and enhance the fidelity of local details in surface reconstruction, and proposes to exploit numerical gradients, replacing conventional analytical computations.
- 2
- Navigational outcomes with a depth-based vision processing algorithm in a second-generation suprachoroidal retinal prosthesis2
Evaluating the effectiveness of a novel depth-based vision processing (VP) method, local background enclosure (LBE), in comparison to the comprehensive VP method, Lanczos2 (L2), in suprachoroidal retinal prosthesis implant recipients during navigational tasks in laboratory and real-world settings found real world outcomes can be improved using an enhanced VP algorithm.
- Open Set Label Shift with Test Time Out-of-Distribution Reference2
Estimators for both source and target open set label distributions are built using a source domain in-distribution (ID) classifier and an ID/OOD classifier to correct the ID classifier trained on the source distribution to the target distribution without retraining.
- SDI-Paste: Synthetic Dynamic Instance Copy-Paste for Video Instance Segmentation2
This paper names their video data augmentation pipeline Synthetic Dynamic Instance Copy-Paste, and test it on the complex task of Video Instance Segmentation which combines detection, segmentation and tracking of object instances across a video sequence.
- 2
- Recursive Training for Zero-Shot Semantic Segmentation2
This work proposes a recursive training scheme to supervise the retraining of a semantic segmentation model for a zero-shot setting using a pseudo-feature representation, and proposes a Zero-Shot Maximum Mean Discrepancy (ZS-MMD) loss that weighs high confidence outputs of the pixel classification layer as a Pseudo-feature representations, and feeds it back to the generator.
- 2
- Learning To Segment Dominant Object Motion From Watching Videos2
Inspired by a layered image representation, a technique to group pixel regions according to their affine parametric motion is introduced, which enables the network to learn segmentation of the dominant foreground object using only RGB image pairs as input for both training and inference.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- Bis(methyltri-o-tolylphosphonium) octaiodide2
This is the first example of a structure containing an [(o-tolyl)(3)PMe](+) cation, and the C(Me)-P-C-C(Me) torsion angles indicate that the configuration is exo(3), which is linked via long-range interactions into infinite polyiodide ribbons.
- 2
- Cochlea Modelling: Clinical Challenges and Tubular Extraction2
The 3-dimensional cochlea representation described herein is the first to explicitly integrate path and cross-sectional shape, specifically directed at addressing clinical outcomes.
- System Results and Case Studies2
This chapter presents the experimental results, which consist of two parts: the major components, the recognition and pose determination, are evaluated, and a number of case studies showing the performance of the integrated robot circumnavigation system.
- 1
- Learning Gaussian Representation for Eye Fixation Prediction1
This work proposes to model the eye fixation map as a mixture of probability distributions, namely a Gaussian Mixture Model, and designs a framework upon some lightweight backbones to achieve real-time fixation prediction.
- Attention Prior for Real Image Restoration.1
A novel single-stage blind real image restoration network (R$^2$Net) is proposed by employing a modular architecture that uses a residual on the residual structure to ease the flow of low-frequency information and apply feature attention to exploit the channel dependencies.
- 1
- 1
- 1
- 1
- 1
- Caricaturing improves face identity recognition in simulated prosthetic vision1
Caricaturing offers practical benefits for face recognition with the bionic eye by enhancing high-level face shape information and improving both the perceptual individuation and recognition of phoshenised faces.
- 1
- 1
- 1
- 1
- Towards a Hazard Perception Assistance System using Visual Motion1
Preliminary work in the application of low- level visual motion cues to identify potential hazards during on-road driving using the use of visual motion as a means of estimating self motion, from which optical flow due to other motions in the scene is identified.
- Real Time Motion Recovery using a Hemispherical Sensor1
A new implementation of a wide angle camera visual motion recovery algorithm suitable for use on a low cost, low power, light-weight wearable sensing device is presented.
- Comparison of Temporal Filters for Optical Flow Estimation in Continuous Mobile Robot Navigation1
Comparisons of three spatio-temporal filters used in the estimation of optical flow for continuous mobile robot navigation indicate that long latencies significantly impede performance for real-time tasks in the control loop.
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- Learned and Hand-crafted Feature Fusion in Unit Ball for 3D Object Classification–
A light-weight experimental architecture for 3D object classification, that operates in B 3, and shows that the optimal view angles for extracting patterns from 3D objects depend on its shape and achieve compelling results with a relatively shallow network, compared to the state-of-the-art.
- –
- How to train your conditional GAN: An approach using geometrically structured latent manifolds.–
This paper argues that the limited diversity of the vanilla cGANs is not due to a lack of capacity, but a result of non-optimal training schemes, and proposes a novel training mechanism that increases both the diversity and the visual quality of the Vanilla cGAN.
- –
- Deep Image Smoothing based on Texture and Structure Guidance.–
This paper builds a texture prediction network that learns from a various of natural textures and combines this with a structure prediction network (SPN) so that the final double-guided filtering network (DGFN) is informed where are the textures to remove andWhere are the structures to preserve.
- –
- –
- –
- –
- –
- –
- –
- –
- Ravenbrook Chart: A New Library for Graph Layout and Visualization–
Ravenbrook Chart, a newly-available library which implements layout and interactive display of large, complex graphs, is now available free of charge as a permissively licensed library.
- –
- AI@NICTA–
Much of the work taking place at NICTA involves various topics in artificial intelligence, and some of the AI work being undertaken is surveyed.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Primate structures in synthetic dynamic active visual saliency–
The structures and functional pathways of the synthetic system form an architecture broadly similar to that observed from neural correlates in primates, and the correspondence to biology has emerged naively, as a result of performance optimisation.
- Edge Segmentation and Matching–
This chapter describes the edge segmentation and matching technique that was developed to support the canonical-views modelling method described in Chapter 4, and was used for initial edge extraction.
- Object Recognition Mobile Robot Guidance–
This chapter describes a method for object recognition and navigation that is designed specifically for robot interaction, based on canonical-views, a novel object representation which realises principles of embodied vision.
- Supporting Navigation Components–
This chapter describes aspects of the robot system that are essential to its operation and describes a path planning approach and the interface with an odometry and obstacle detection system.
- Embodied Vision For Mobile Robots–
This chapter discusses philosophical aspects of what is required to apply of high-level vision when the perceiver has a real body, i.e., an autonomous mobile robot and proposes a framework of embodied categorisation for high- level vision and interaction for mobile robots.
- Related Systems and Ideas–
This chapter presents some basic vision concepts and techniques to facilitate discussion of vision-guidance for mobile robots, and describes some of the most effective for vision-guided robot navigation systems.
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.