Publications (122)
Learning Positional Embeddings for Coordinate-MLPs
Sameera Ramasinghe, Simon Lucey
We propose a novel method to enhance the performance of coordinate-MLPs by learning instance-specific positional embeddings. End-to-end optimization of positional embedding paramet…
Rethinking Reprojection: Closing the Loop for Pose-aware ShapeReconstruction from a Single Image
Rui Zhu, Hamed Kiani Galoogahi, Chaoyang Wang +1
An emerging problem in computer vision is the reconstruction of 3D shape and pose of an object from a single image. Hitherto, the problem has been addressed through the application…
Parameter-Efficient Fine-Tuning with Learnable Rank
Arpit Garg, Simon Lucey, Hemanth Saratchandran
Low-Rank Adaptation (LoRA) is a popular parameter-efficient fine-tuning (PEFT) method that restricts weight updates to low-rank adapters, introducing a fixed low-rank inductive bia…
Optimization Methods for Convolutional Sparse Coding
Hilton Bristow, Simon Lucey
Sparse and convolutional constraints form a natural prior for many optimization problems that arise from physical processes. Detecting motifs in speech and musical passages, super-…
Learning detectors quickly using structured covariance matrices
Jack Valmadre, Sridha Sridharan, Simon Lucey
Computer vision is increasingly becoming interested in the rapid estimation of object detectors. Canonical hard negative mining strategies are slow as they require multiple passes…
Deep Non-Rigid Structure from Motion with Missing Data
Chen Kong, Simon Lucey
Non-Rigid Structure from Motion (NRSfM) refers to the problem of reconstructing cameras and the 3D point cloud of a non-rigid object from an ensemble of images with 2D corresponden…
Convolutional Initialization for Data-Efficient Vision Transformers
Jianqiao Zheng, Xueqian Li, Simon Lucey
Training vision transformer networks on small datasets poses challenges. In contrast, convolutional neural networks (CNNs) can achieve state-of-the-art performance by leveraging th…
Image2Mesh: A Learning Framework for Single Image 3D Reconstruction
Jhony K. Pontes, Chen Kong, Sridha Sridharan +3
One challenge that remains open in 3D deep learning is how to efficiently represent 3D data to feed deep networks. Recent works have relied on volumetric or point cloud representat…
Learning Background-Aware Correlation Filters for Visual Tracking
Hamed Kiani Galoogahi, Ashton Fagg, Simon Lucey
Correlation Filters (CFs) have recently demonstrated excellent performance in terms of rapidly tracking objects under challenging photometric and geometric variations. The strength…
Neural Trajectory Fields for Dynamic Novel View Synthesis
Chaoyang Wang, Ben Eckart, Simon Lucey +1
Recent approaches to render photorealistic views from a limited set of photographs have pushed the boundaries of our interactions with pictures of static scenes. The ability to rec…
On progressive sharpening, flat minima and generalisation
Lachlan Ewen MacDonald, Jack Valmadre, Simon Lucey
We present a new approach to understanding the relationship between loss curvature and input-output model behaviour in deep learning. Specifically, we use existing empirical analys…
Enhancing Transformers Through Conditioned Embedded Tokens
Hemanth Saratchandran, Simon Lucey
Transformers have transformed modern machine learning, driving breakthroughs in computer vision, natural language processing, and robotics. At the core of their success lies the at…
The Conditional Lucas & Kanade Algorithm
Chen-Hsuan Lin, Rui Zhu, Simon Lucey
The Lucas & Kanade (LK) algorithm is the method of choice for efficient dense image and object alignment. The approach is efficient as it attempts to model the connection between a…
Multi-Body Neural Scene Flow
Kavisha Vidanapathirana, Shin-Fang Chng, Xueqian Li +1
The test-time optimization of scene flow - using a coordinate network as a neural prior - has gained popularity due to its simplicity, lack of dataset bias, and state-of-the-art pe…
How You Start Matters for Generalization
Sameera Ramasinghe, Lachlan MacDonald, Moshiur Farazi +2
Characterizing the remarkable generalization properties of over-parameterized neural networks remains an open problem. In this paper, we promote a shift of focus towards initializa…
Distill Knowledge from NRSfM for Weakly Supervised 3D Pose Learning
Chaoyang Wang, Chen Kong, Simon Lucey
We propose to learn a 3D pose estimator by distilling knowledge from Non-Rigid Structure from Motion (NRSfM). Our method uses solely 2D landmark annotations. No 3D data, multi-view…
Aligning Across Large Gaps in Time
Hunter Goforth, Simon Lucey
We present a method of temporally-invariant image registration for outdoor scenes, with invariance across time of day, across seasonal variations, and across decade-long periods, f…
3D-LFM: Lifting Foundation Model
Mosam Dabhi, Laszlo A. Jeni, Simon Lucey
The lifting of 3D structure and camera from 2D landmarks is at the cornerstone of the entire discipline of computer vision. Traditional methods have been confined to specific rigid…
Long-term Visual Map Sparsification with Heterogeneous GNN
Ming-Fang Chang, Yipu Zhao, Rajvi Shah +3
We address the problem of map sparsification for long-term visual localization. For map sparsification, a commonly employed assumption is that the pre-build map and the later captu…
SeMoLi: What Moves Together Belongs Together
Jenny Seidenschwarz, Aljoša Ošep, Francesco Ferroni +2
We tackle semi-supervised object detection based on motion cues. Recent results suggest that heuristic-based clustering methods in conjunction with object trackers can be used to p…
Re-Evaluating LiDAR Scene Flow for Autonomous Driving
Nathaniel Chodosh, Deva Ramanan, Simon Lucey
Popular benchmarks for self-supervised LiDAR scene flow (stereoKITTI, and FlyingThings3D) have unrealistic rates of dynamic motion, unrealistic correspondences, and unrealistic sam…
Invertible Neural Warp for NeRF
Shin-Fang Chng, Ravi Garg, Hemanth Saratchandran +1
This paper tackles the simultaneous optimization of pose and Neural Radiance Fields (NeRF). Departing from the conventional practice of using explicit global representations for ca…
Architectural Adversarial Robustness: The Case for Deep Pursuit
George Cazenavette, Calvin Murdock, Simon Lucey
Despite their unmatched performance, deep neural networks remain susceptible to targeted attacks by nearly imperceptible levels of adversarial noise. While the underlying cause of…
SDF-SRN: Learning Signed Distance 3D Object Reconstruction from Static Images
Chen-Hsuan Lin, Chaoyang Wang, Simon Lucey
Dense 3D object reconstruction from a single image has recently witnessed remarkable advances, but supervising neural networks with ground-truth 3D shapes is impractical due to the…
Curvature-Aware Training for Coordinate Networks
Hemanth Saratchandran, Shin-Fang Chng, Sameera Ramasinghe +2
Coordinate networks are widely used in computer vision due to their ability to represent signals as compressed, continuous entities. However, training these networks with first-ord…
Compact Model Representation for 3D Reconstruction
Jhony K. Pontes, Chen Kong, Anders Eriksson +3
3D reconstruction from 2D images is a central problem in computer vision. Recent works have been focusing on reconstruction directly from a single image. It is well known however t…
Take it in your stride: Do we need striding in CNNs?
Chen Kong, Simon Lucey
Since their inception, CNNs have utilized some type of striding operator to reduce the overlap of receptive fields and spatial dimensions. Although having clear heuristic motivatio…
Deep Convolutional Compressed Sensing for LiDAR Depth Completion
Nathaniel Chodosh, Chaoyang Wang, Simon Lucey
In this paper we consider the problem of estimating a dense depth map from a set of sparse LiDAR points. We use techniques from compressed sensing and the recently developed Altern…
Bit-Planes: Dense Subpixel Alignment of Binary Descriptors
Hatem Alismail, Brett Browning, Simon Lucey
Binary descriptors have been instrumental in the recent evolution of computationally efficient sparse image alignment algorithms. Increasingly, however, the vision community is int…
ST-GAN: Spatial Transformer Generative Adversarial Networks for Image Compositing
Chen-Hsuan Lin, Ersin Yumer, Oliver Wang +2
We address the problem of finding realistic geometric corrections to a foreground object such that it appears natural when composited into a background image. To achieve this, we p…
Fast Kernel Scene Flow
Xueqian Li, Simon Lucey
In contrast to current state-of-the-art methods, such as NSFP [25], which employ deep implicit neural functions for modeling scene flow, we present a novel approach that utilizes c…
GARF: Gaussian Activated Radiance Fields for High Fidelity Reconstruction and Pose Estimation
Shin-Fang Chng, Sameera Ramasinghe, Jamie Sherrah +1
Despite Neural Radiance Fields (NeRF) showing compelling results in photorealistic novel views synthesis of real-world scenes, most existing approaches require accurate prior camer…
Deep Interpretable Non-Rigid Structure from Motion
Chen Kong, Simon Lucey
All current non-rigid structure from motion (NRSfM) algorithms are limited with respect to: (i) the number of images, and (ii) the type of shape variability they can handle. This h…
Leaner Transformers: More Heads, Less Depth
Hemanth Saratchandran, Damien Teney, Simon Lucey
Transformers have reshaped machine learning by utilizing attention mechanisms to capture complex patterns in large datasets, leading to significant improvements in performance. Thi…
Gradient Descent as a Shrinkage Operator for Spectral Bias
Simon Lucey
We generalize the connection between activation function and spline regression/smoothing and characterize how this choice may influence spectral bias within a 1D shallow network. W…
Regression-Based Image Alignment for General Object Categories
Hilton Bristow, Simon Lucey
Gradient-descent methods have exhibited fast and reliable performance for image alignment in the facial domain, but have largely been ignored by the broader vision community. They…
Object Agnostic 3D Lifting in Space and Time
Christopher Fusco, Shin-Fang Ch'ng, Mosam Dabhi +1
We present a spatio-temporal perspective on category-agnostic 3D lifting of 2D keypoints over a temporal sequence. Our approach differs from existing state-of-the-art methods that…
The Quantization Benefits of Residual-Free Transformers
Yiping Ji, Mahalakshmi Sabanayagam, Peyman Moghadam +2
Large-scale transformer training and deployment are increasingly constrained by the transfer of activations, gradients, and optimizer states across accelerators. Low-bit quantizati…
Can Transformers Really Do It All? On the Compatibility of Inductive Biases Across Tasks
Damien Teney, Liangze Jiang, Hemanth Saratchandran +1
Transformers are remarkably versatile and their design is largely consistent across a variety of applications. But are they optimal for any given task or dataset? The answer may be…
PAUL: Procrustean Autoencoder for Unsupervised Lifting
Chaoyang Wang, Simon Lucey
Recent success in casting Non-rigid Structure from Motion (NRSfM) as an unsupervised deep learning problem has raised fundamental questions about what novelty in NRSfM prior could…
Structured Initialization for Attention in Vision Transformers
Jianqiao Zheng, Xueqian Li, Simon Lucey
The training of vision transformer (ViT) networks on small-scale datasets poses a significant challenge. By contrast, convolutional neural networks (CNNs) have an architectural ind…
Efficient Learning With Sine-Activated Low-rank Matrices
Yiping Ji, Hemanth Saratchandran, Cameron Gordon +2
Low-rank decomposition has emerged as a vital tool for enhancing parameter efficiency in neural network architectures, gaining traction across diverse applications in machine learn…
Web Stereo Video Supervision for Depth Prediction from Dynamic Scenes
Chaoyang Wang, Simon Lucey, Federico Perazzi +1
We present a fully data-driven method to compute depth from diverse monocular video sequences that contain large amounts of non-rigid objects, e.g., people. In order to learn recon…
Structured Initialization for Vision Transformers
Jianqiao Zheng, Xueqian Li, Hemanth Saratchandran +1
Convolutional Neural Networks (CNNs) inherently encode strong inductive biases, enabling effective generalization on small-scale datasets. In this paper, we propose integrating thi…
3D Gaussian Point Encoders
Jim James, Ben Wilson, Simon Lucey +1
In this work, we introduce the 3D Gaussian Point Encoder, an explicit per-point embedding built on mixtures of learned 3D Gaussians. This explicit geometric representation for 3D r…
Weight Conditioning for Smooth Optimization of Neural Networks
Hemanth Saratchandran, Thomas X. Wang, Simon Lucey
In this article, we introduce a novel normalization technique for neural network weight matrices, which we term weight conditioning. This approach aims to narrow the gap between th…
Learning Efficient Point Cloud Generation for Dense 3D Object Reconstruction
Chen-Hsuan Lin, Chen Kong, Simon Lucey
Conventional methods of 3D object generative modeling learn volumetric predictions using deep networks with 3D convolutional operations, which are direct analogies to classical 2D…
A Sampling Theory Perspective on Activations for Implicit Neural Representations
Hemanth Saratchandran, Sameera Ramasinghe, Violetta Shevchenko +2
Implicit Neural Representations (INRs) have gained popularity for encoding signals as compact, differentiable entities. While commonly using techniques like Fourier positional enco…
From Tables to Signals: Revealing Spectral Adaptivity in TabPFN
Jianqiao Zheng, Cameron Gordon, Yiping Ji +2
Task-agnostic tabular foundation models such as TabPFN have achieved impressive performance on tabular learning tasks, yet the origins of their inductive biases remain poorly under…
BARF: Bundle-Adjusting Neural Radiance Fields
Chen-Hsuan Lin, Wei-Chiu Ma, Antonio Torralba +1
Neural Radiance Fields (NeRF) have recently gained a surge of interest within the computer vision community for its power to synthesize photorealistic novel views of real-world sce…
On the effectiveness of neural priors in modeling dynamical systems
Sameera Ramasinghe, Hemanth Saratchandran, Violetta Shevchenko +1
Modelling dynamical systems is an integral component for understanding the natural world. To this end, neural networks are becoming an increasingly popular candidate owing to their…
On the Bias Against Inductive Biases
George Cazenavette, Simon Lucey
Borrowing from the transformer models that revolutionized the field of natural language processing, self-supervised feature learning for visual tasks has also seen state-of-the-art…
One Framework to Register Them All: PointNet Encoding for Point Cloud Alignment
Vinit Sarode, Xueqian Li, Hunter Goforth +5
PointNet has recently emerged as a popular representation for unstructured point cloud data, allowing application of deep learning to tasks such as object detection, segmentation a…
Cutting the Skip: Training Residual-Free Transformers
Yiping Ji, James Martens, Jianqiao Zheng +5
Transformers have achieved remarkable success across a wide range of applications, a feat often attributed to their scalability. Yet training them without skip (residual) connectio…
G3Splat: Geometrically Consistent Generalizable Gaussian Splatting
Mehdi Hosseinzadeh, Shin-Fang Chng, Yi Xu +3
3D Gaussians have become a powerful scene representation for real-time splatting and high-quality novel-view synthesis. This has motivated generalizable splatting -- methods that a…
Robust Physical Adversarial Patches Using Dynamically Optimized Clusters
Harrison Bagley, Will Meakin, Simon Lucey +2
Physical adversarial attacks on deep learning systems is concerning due to the ease of deploying such attacks, usually by placing an adversarial patch in a scene to manipulate the…
When to Use Convolutional Neural Networks for Inverse Problems
Nathaniel Chodosh, Simon Lucey
Reconstruction tasks in computer vision aim fundamentally to recover an undetermined signal from a set of noisy measurements. Examples include super-resolution, image denoising, an…
SineProject: Machine Unlearning for Stable Vision Language Alignment
Arpit Garg, Hemanth Saratchandran, Simon Lucey
Multimodal Large Language Models (MLLMs) increasingly need to forget specific knowledge such as unsafe or private information without requiring full retraining. However, existing u…
Photometric Mesh Optimization for Video-Aligned 3D Object Reconstruction
Chen-Hsuan Lin, Oliver Wang, Bryan C. Russell +4
In this paper, we address the problem of 3D object mesh reconstruction from RGB videos. Our approach combines the best of multi-view geometric and data-driven methods for 3D recons…
Deep-LK for Efficient Adaptive Object Tracking
Chaoyang Wang, Hamed Kiani Galoogahi, Chen-Hsuan Lin +1
In this paper we present a new approach for efficient regression based object tracking which we refer to as Deep- LK. Our approach is closely related to the Generic Object Tracking…
Beyond Periodicity: Towards a Unifying Framework for Activations in Coordinate-MLPs
Sameera Ramasinghe, Simon Lucey
Coordinate-MLPs are emerging as an effective tool for modeling multidimensional continuous signals, overcoming many drawbacks associated with discrete grid-based approximations. Ho…
MaskNet: A Fully-Convolutional Network to Estimate Inlier Points
Vinit Sarode, Animesh Dhagat, Rangaprasad Arun Srivatsan +3
Point clouds have grown in importance in the way computers perceive the world. From LIDAR sensors in autonomous cars and drones to the time of flight and stereo vision systems in o…
D'OH: Decoder-Only Random Hypernetworks for Implicit Neural Representations
Cameron Gordon, Lachlan Ewen MacDonald, Hemanth Saratchandran +1
Deep implicit functions have been found to be an effective tool for efficiently encoding all manner of natural signals. Their attractiveness stems from their ability to compactly r…
Stable Forgetting: Bounded Parameter-Efficient Unlearning in Foundation Models
Arpit Garg, Hemanth Saratchandran, Ravi Garg +1
Machine unlearning in foundation models (e.g., language and vision transformers) is essential for privacy and safety; however, existing approaches are unstable and unreliable. A wi…
Rethinking Attention: Polynomial Alternatives to Softmax in Transformers
Hemanth Saratchandran, Jianqiao Zheng, Yiping Ji +2
This paper questions whether the strong performance of softmax attention in transformers stems from producing a probability distribution over inputs. Instead, we argue that softmax…
On Regularizing Coordinate-MLPs
Sameera Ramasinghe, Lachlan MacDonald, Simon Lucey
We show that typical implicit regularization assumptions for deep neural networks (for regression) do not hold for coordinate-MLPs, a family of MLPs that are now ubiquitous in comp…
Reframing Neural Networks: Deep Structure in Overcomplete Representations
Calvin Murdock, George Cazenavette, Simon Lucey
In comparison to classical shallow representation learning techniques, deep neural networks have achieved superior performance in nearly every application benchmark. But despite th…
Architectural Strategies for the optimization of Physics-Informed Neural Networks
Hemanth Saratchandran, Shin-Fang Chng, Simon Lucey
Physics-informed neural networks (PINNs) offer a promising avenue for tackling both forward and inverse problems in partial differential equations (PDEs) by incorporating deep lear…
Semantic Photometric Bundle Adjustment on Natural Sequences
Rui Zhu, Chaoyang Wang, Chen-Hsuan Lin +2
The problem of obtaining dense reconstruction of an object in a natural sequence of images has been long studied in computer vision. Classically this problem has been solved throug…
Learning Depth from Monocular Videos using Direct Methods
Chaoyang Wang, Jose Miguel Buenaposada, Rui Zhu +1
The ability to predict depth from a single image - using recent advances in CNNs - is of increasing interest to the vision community. Unsupervised strategies to learning are partic…
Always Skip Attention
Yiping Ji, Hemanth Saratchandran, Peyman Moghadam +1
We highlight a curious empirical result within modern Vision Transformers (ViTs). Specifically, self-attention catastrophically fails to train unless it is used in conjunction with…
Flow supervision for Deformable NeRF
Chaoyang Wang, Lachlan Ewen MacDonald, Laszlo A. Jeni +1
In this paper we present a new method for deformable NeRF that can directly use optical flow as supervision. We overcome the major challenge with respect to the computationally ine…
Correlation Filters with Limited Boundaries
Hamed Kiani Galoogahi, Terence Sim, Simon Lucey
Correlation filters take advantage of specific properties in the Fourier domain allowing them to be estimated efficiently: O(NDlogD) in the frequency domain, versus O(D^3 + ND^2) s…
Evading the Simplicity Bias: Training a Diverse Set of Models Discovers Solutions with Superior OOD Generalization
Damien Teney, Ehsan Abbasnejad, Simon Lucey +1
Neural networks trained with SGD were recently shown to rely preferentially on linearly-predictive features and can ignore complex, equally-predictive ones. This simplicity bias ca…
Joint Max Margin and Semantic Features for Continuous Event Detection in Complex Scenes
Iman Abbasnejad, Sridha Sridharan, Simon Denman +2
In this paper the problem of complex event detection in the continuous domain (i.e. events with unknown starting and ending locations) is addressed. Existing event detection method…
Argoverse: 3D Tracking and Forecasting with Rich Maps
Ming-Fang Chang, John Lambert, Patsorn Sangkloy +8
We present Argoverse -- two datasets designed to support autonomous vehicle machine learning tasks such as 3D tracking and motion forecasting. Argoverse was collected by a fleet of…
Fast Neural Scene Flow
Xueqian Li, Jianqiao Zheng, Francesco Ferroni +2
Neural Scene Flow Prior (NSFP) is of significant interest to the vision community due to its inherent robustness to out-of-distribution (OOD) effects and its ability to deal with d…
Scene Flow from Point Clouds with or without Learning
Jhony Kaesemodel Pontes, James Hays, Simon Lucey
Scene flow is the three-dimensional (3D) motion field of a scene. It provides information about the spatial arrangement and rate of change of objects in dynamic environments. Curre…
Direct Visual Odometry using Bit-Planes
Hatem Alismail, Brett Browning, Simon Lucey
Feature descriptors, such as SIFT and ORB, are well-known for their robustness to illumination changes, which has made them popular for feature-based VSLAM\@. However, in degraded…
Deep Non-Rigid Structure from Motion
Chen Kong, Simon Lucey
Current non-rigid structure from motion (NRSfM) algorithms are mainly limited with respect to: (i) the number of images, and (ii) the type of shape variability they can handle. Thi…
Learning Policies for Adaptive Tracking with Deep Feature Cascades
Chen Huang, Simon Lucey, Deva Ramanan
Visual object tracking is a fundamental and time-critical vision task. Recent years have seen many shallow tracking methods based on real-time pixel-based correlation filters, as w…
Spectral Conditioning of Attention Improves Transformer Performance
Hemanth Saratchandran, Simon Lucey
We present a theoretical analysis of the Jacobian of an attention block within a transformer, showing that it is governed by the query, key, and value projections that define the a…
Inverse Compositional Spatial Transformer Networks
Chen-Hsuan Lin, Simon Lucey
In this paper, we establish a theoretical connection between the classical Lucas & Kanade (LK) algorithm and the emerging topic of Spatial Transformer Networks (STNs). STNs are of…
Robust Point Cloud Processing through Positional Embedding
Jianqiao Zheng, Xueqian Li, Sameera Ramasinghe +1
End-to-end trained per-point embeddings are an essential ingredient of any state-of-the-art 3D point cloud processing such as detection or alignment. Methods like PointNet, or the…
Dense Semantic Correspondence where Every Pixel is a Classifier
Hilton Bristow, Jack Valmadre, Simon Lucey
Determining dense semantic correspondences across objects and scenes is a difficult problem that underpins many higher-level computer vision algorithms. Unlike canonical dense corr…
Neural Scene Flow Prior
Xueqian Li, Jhony Kaesemodel Pontes, Simon Lucey
Before the deep learning revolution, many perception algorithms were based on runtime optimization in conjunction with a strong prior/regularization penalty. A prime example of thi…
PointNetLK Revisited
Xueqian Li, Jhony Kaesemodel Pontes, Simon Lucey
We address the generalization ability of recent learning-based point cloud registration methods. Despite their success, these approaches tend to have poor performance when applied…
Joint Pose and Shape Estimation of Vehicles from LiDAR Data
Hunter Goforth, Xiaoyan Hu, Michael Happold +1
We address the problem of estimating the pose and shape of vehicles from LiDAR scans, a common problem faced by the autonomous vehicle community. Recent work has tended to address…
PCRNet: Point Cloud Registration Network using PointNet Encoding
Vinit Sarode, Xueqian Li, Hunter Goforth +4
PointNet has recently emerged as a popular representation for unstructured point cloud data, allowing application of deep learning to tasks such as object detection, segmentation a…
Learning Temporal Alignment Uncertainty for Efficient Event Detection
Iman Abbasnejad, Sridha Sridharan, Simon Denman +2
In this paper we tackle the problem of efficient video event detection. We argue that linear detection functions should be preferred in this regard due to their scalability and eff…
Rethinking the Role of Spatial Mixing
George Cazenavette, Joel Julin, Simon Lucey
Until quite recently, the backbone of nearly every state-of-the-art computer vision model has been the 2D convolution. At its core, a 2D convolution simultaneously mixes informatio…
Rethinking Positional Encoding
Jianqiao Zheng, Sameera Ramasinghe, Simon Lucey
It is well noted that coordinate based MLPs benefit -- in terms of preserving high-frequency information -- through the encoding of coordinate positions as an array of Fourier feat…
Dataless Model Selection with the Deep Frame Potential
Calvin Murdock, Simon Lucey
Choosing a deep neural network architecture is a fundamental problem in applications that require balancing performance and parameter efficiency. Standard approaches rely on ad-hoc…
Proxy Templates for Inverse Compositional Photometric Bundle Adjustment
Christopher Ham, Simon Lucey, Surya Singh
Recent advances in 3D vision have demonstrated the strengths of photometric bundle adjustment. By directly minimizing reprojected pixel errors, instead of geometric reprojection er…
DARB-Splatting: Generalizing Splatting with Decaying Anisotropic Radial Basis Functions
Hashiru Pramuditha, Vinasirajan Viruthshaan, Vishagar Arunan +4
Splatting-based 3D reconstruction methods have gained popularity with the advent of 3D Gaussian Splatting, efficiently synthesizing high-quality novel views. These methods commonly…
On skip connections and normalisation layers in deep optimisation
Lachlan Ewen MacDonald, Jack Valmadre, Hemanth Saratchandran +1
We introduce a general theoretical framework, designed for the study of gradient optimisation of deep neural networks, that encompasses ubiquitous architecture choices including ba…
High Fidelity 3D Reconstructions with Limited Physical Views
Mosam Dabhi, Chaoyang Wang, Kunal Saluja +3
Multi-view triangulation is the gold standard for 3D reconstruction from 2D correspondences given known calibration and sufficient views. However in practice, expensive multi-view…
Mask the Target: A Plug-and-Play Regularizer Against LoRA Forgetting
Runze Xu, Arpit Garg, Hemanth Saratchandran +1
Low-Rank Adaptation (LoRA) has become one of the most widely used fine-tuning mechanisms for adapting large language models to new domains, tasks, and users. Yet adaptation perform…
Preconditioners for the Stochastic Training of Neural Fields
Shin-Fang Chng, Hemanth Saratchandran, Simon Lucey
Neural fields encode continuous multidimensional signals as neural networks, enabling diverse applications in computer vision, robotics, and geometry. While Adam is effective for s…
Memory Efficient Tabular Foundation Models
Shuting Luo, Monika Mikhail Kanaan, Cameron Gordon +2
The paper studies how to reduce the memory footprint of tabular foundation models like TabPFN using compression techniques, achieving up to 7.6× memory savings with little performa…