arXiv subjects

Michael Green

Publications and source records attributed to Michael Green.

13 recordsLinked to original sources

arXiv2026

WarpHammer: Densifying Scene Warps with 3D Object Priors for Extreme View Synthesis

Projection-conditioned novel view synthesis (NVS) warps an explicit 3D reconstruction of the input view into the target camera and conditions a generator on the warped rendering. This works well for small viewpoint changes but degrades sharply under large orbital motion: the warp becomes sparse around the orbited object, where hidden surfaces dominate the new view and mirror-like artifacts emerge, causing the generator to lose both pixel content and the implicit camera cue carried by the warp. We introduce WarpHammer, a training-free framework that resolves this failure mode by augmenting the warped scene with an explicit 3D reconstruction of the object obtained from a native 3D generative prior (e.g., SAM3D). The reconstructed object adds missing foreground surfaces and occludes background points that should no longer be visible, restoring both appearance and camera cues without fine-tuning the base model. The same explicit object representation further unlocks a capability current NVS pipelines do not support: incorporating auxiliary views of the object from sources outside the target scene, for example, a casual snapshot of a car paired with a manufacturer studio shot of the same model. We process the reference and auxiliary images jointly with a pretrained multi-view geometry foundation model, which predicts a unified point cloud that we fuse into the 3D object reconstruction. This yields substantially more faithful geometry than single-image reconstruction, without requiring user-provided camera poses for the auxiliary views. On five benchmarks, WarpHammer produces stable novel views at viewpoint deviations where strong baselines collapse, and is the first scene-level NVS method that can naturally fuse auxiliary, pose-unknown object views from an external source.

cs.CV↗

arXiv2026

VidMsg: A Benchmark for Implicit Message Inference in Short Videos

Understanding short online videos involves more than identifying visible objects and actions; video makers often include an underlying message or purpose in the clip. We introduce VidMsg, a benchmark for evaluating implicit message understanding in short, internet-native video clips. VidMsg contains 400 YouTube-derived clips across 9 practical topic areas and 52 fine-grained target messages, covering domains such as career and finance, education, health and well-being, culture, safety, sustainability, and lifestyle. VidMsg is constructed through a message-first pipeline: an LLM first translates target messages into indirect search scenarios, which are used to retrieve candidate clips. Human annotators then retain clips that convey the intended message without being overly explicit. VidMsg is designed primarily for bidirectional message-clip retrieval for scalable applications such as video search and recommendation, where systems must capture holistic video understanding. In addition to retrieval, VidMsg includes a diagnostic multiple-choice QA benchmark, where models select the intended message of a clip from semantically related alternatives. Experiments with contemporary video-language and retrieval models show that strong models often fail on VidMsg, because the task requires pragmatic inference, integration of contextual cues, and discrimination among semantically close messages. We also introduce VidVec-Msg, a baseline method that improves message-oriented retrieval while leaving substantial headroom for future work.

cs.CV↗

arXiv2026

Fast Autoregressive Video Diffusion and World Models with Temporal Cache Compression and Sparse Attention

Autoregressive video diffusion models enable streaming generation, opening the door to long-form synthesis, video world models, and interactive neural game engines. However, their core attention layers become a major bottleneck at inference time: as generation progresses, the KV cache grows, causing both increasing latency and escalating GPU memory, which in turn restricts usable temporal context and harms long-range consistency. In this work, we study redundancy in autoregressive video diffusion and identify three persistent sources: near-duplicate cached keys across frames, slowly evolving (largely semantic) queries/keys that make many attention computations redundant, and cross-attention over long prompts where only a small subset of tokens matters per frame. Building on these observations, we propose a unified, training-free attention framework (FAST-AR) for FAST-AutoRegressive diffusion, consisting of three components: TempCache compresses the KV cache via temporal correspondence to bound cache growth; AnnCA accelerates cross-attention by selecting frame-relevant prompt tokens using fast approximate nearest neighbor (ANN) matching; and AnnSA sparsifies self-attention by restricting each query to semantically matched keys, also using a lightweight ANN. Together, these modules reduce attention, compute, and memory and are compatible with existing autoregressive diffusion backbones and world models. Experiments demonstrate up to x5 - x10 end-to-end speedups while preserving near-identical visual quality and, crucially, maintaining stable throughput and nearly constant peak GPU memory usage over long rollouts, where prior methods progressively slow down and suffer from increasing memory usage.

cs.CV↗

arXiv2025

MIP-Based Tumor Segmentation: A Radiologist-Inspired Approach

PET/CT imaging is the gold standard for tumor detection, offering high accuracy in identifying local and metastatic lesions. Radiologists often begin assessment with rotational Multi-Angle Maximum Intensity Projections (MIPs) from PET, confirming findings with volumetric slices. This workflow is time-consuming, especially in metastatic cases. Despite their clinical utility, MIPs are underutilized in automated tumor segmentation, where 3D volumetric data remains the norm. We propose an alternative approach that trains segmentation models directly on MIPs, bypassing the need to segment 3D volumes and then project. This better aligns the model with its target domain and yields substantial gains in computational efficiency and training time. We also introduce a novel occlusion correction method that restores MIP annotations occluded by high-intensity structures, improving segmentation. Using the autoPET 2022 Grand Challenge dataset, we evaluate our method against standard 3D pipelines in terms of performance and training/computation efficiency for segmentation and classification, and analyze how MIP count affects segmentation. Our MIP-based approach achieves segmentation performance on par with 3D (<=1% Dice difference, 26.7% better Hausdorff Distance), while reducing training time (convergence time) by 55.8-75.8%, energy per epoch by 71.7-76%, and TFLOPs by two orders of magnitude, highlighting its scalability for clinical use. For classification, using 16 MIPs only as input, we surpass 3D performance while reducing training time by over 10x and energy consumption per epoch by 93.35%. Our analysis of the impact of MIP count on segmentation identified 48 views as optimal, offering the best trade-off between performance and efficiency.

eess.IV↗

arXiv2025

Find your Needle: Small Object Image Retrieval via Multi-Object Attention Optimization

We address the challenge of Small Object Image Retrieval (SoIR), where the goal is to retrieve images containing a specific small object, in a cluttered scene. The key challenge in this setting is constructing a single image descriptor, for scalable and efficient search, that effectively represents all objects in the image. In this paper, we first analyze the limitations of existing methods on this challenging task and then introduce new benchmarks to support SoIR evaluation. Next, we introduce Multi-object Attention Optimization (MaO), a novel retrieval framework which incorporates a dedicated multi-object pre-training phase. This is followed by a refinement process that leverages attention-based feature extraction with object masks, integrating them into a single unified image descriptor. Our MaO approach significantly outperforms existing retrieval methods and strong baselines, achieving notable improvements in both zero-shot and lightweight multi-object fine-tuning. We hope this work will lay the groundwork and inspire further research to enhance retrieval performance for this highly practical task. Code and Data are available on our project page: $\href{https://pihash2k.github.io/findyourneedle.github.io}{https://pihash2k.github.io/findyourneedle.github.io}$.

cs.CV↗

arXiv2024

EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition

The task of Visual Place Recognition (VPR) is to predict the location of a query image from a database of geo-tagged images. Recent studies in VPR have highlighted the significant advantage of employing pre-trained foundation models like DINOv2 for the VPR task. However, these models are often deemed inadequate for VPR without further fine-tuning on VPR-specific data. In this paper, we present an effective approach to harness the potential of a foundation model for VPR. We show that features extracted from self-attention layers can act as a powerful re-ranker for VPR, even in a zero-shot setting. Our method not only outperforms previous zero-shot approaches but also introduces results competitive with several supervised methods. We then show that a single-stage approach utilizing internal ViT layers for pooling can produce global features that achieve state-of-the-art performance, with impressive feature compactness down to 128D. Moreover, integrating our local foundation features for re-ranking further widens this performance gap. Our method also demonstrates exceptional robustness and generalization, setting new state-of-the-art performance, while handling challenging conditions such as occlusion, day-night transitions, and seasonal variations.

cs.CV↗

arXiv2022

Automatic Error Detection in Integrated Circuits Image Segmentation: A Data-driven Approach

Due to the complicated nanoscale structures of current integrated circuits(IC) builds and low error tolerance of IC image segmentation tasks, most existing automated IC image segmentation approaches require human experts for visual inspection to ensure correctness, which is one of the major bottlenecks in large-scale industrial applications. In this paper, we present the first data-driven automatic error detection approach targeting two types of IC segmentation errors: wire errors and via errors. On an IC image dataset collected from real industry, we demonstrate that, by adapting existing CNN-based approaches of image classification and image translation with additional pre-processing and post-processing techniques, we are able to achieve recall/precision of 0.92/0.93 in wire error detection and 0.96/0.90 in via error detection, respectively.

cs.CV↗

arXiv2022

Borch: A Deep Universal Probabilistic Programming Language

Ever since the Multilayered Perceptron was first introduced the connectionist community has struggled with the concept of uncertainty and how this could be represented in these types of models. This past decade has seen a lot of effort in trying to join the principled approach of probabilistic modeling with the scalable nature of deep neural networks. While the theoretical benefits of this consolidation are clear, there are also several important practical aspects of these endeavors; namely to force the models we create to represent, learn, and report uncertainty in every prediction that is made. Many of these efforts have been based on extending existing frameworks with additional structures. We present Borch, a scalable deep universal probabilistic programming language, built on top of PyTorch. The code is available for download and use in our repository https://gitlab.com/desupervised/borch.

cs.AI↗

arXiv2022

Making a (Counterfactual) Difference One Rationale at a Time

Rationales, snippets of extracted text that explain an inference, have emerged as a popular framework for interpretable natural language processing (NLP). Rationale models typically consist of two cooperating modules: a selector and a classifier with the goal of maximizing the mutual information (MMI) between the "selected" text and the document label. Despite their promises, MMI-based methods often pick up on spurious text patterns and result in models with nonsensical behaviors. In this work, we investigate whether counterfactual data augmentation (CDA), without human assistance, can improve the performance of the selector by lowering the mutual information between spurious signals and the document label. Our counterfactuals are produced in an unsupervised fashion using class-dependent generative models. From an information theoretic lens, we derive properties of the unaugmented dataset for which our CDA approach would succeed. The effectiveness of CDA is empirically evaluated by comparing against several baselines including an improved MMI-based rationale schema on two multi aspect datasets. Our results show that CDA produces rationales that better capture the signal of interest.

cs.CL↗

arXiv2018

Can we leverage rating patterns from traditional users to enhance recommendations for children?

Recommender algorithms performance is often associated with the availability of sufficient historical rating data. Unfortunately, when it comes to children, this data is seldom available. In this paper, we report on an initial analysis conducted to examine the degree to which data about traditional users, i.e., adults, can be leveraged to enhance the recommendation process for children.

cs.IR↗

arXiv2008

The Highly Damped Quasinormal Modes of Extremal Reissner-Nordström and Reissner-Nordström-de Sitter Black Holes

We analyze in detail the highly damped quasinormal modes of $D$-dimensional extremal Reissner-Nordstr$\ddot{\rm{o}}$m and Reissner-Nordstr$\ddot{\rm{o}}$m-de Sitter black holes. We only consider the extremal case where the event horizon and the Cauchy inner horizon coincide. We show that, even though the topology of the Stokes/anti-Stokes lines in the extremal case is different than the non-extremal case, the highly damped quasinormal mode frequencies of extremal black holes match exactly with the extremal limit of the non-extremal black hole quasinormal mode frequencies.

gr-qc↗

arXiv2002

Metric Coordinate Systems

Coordinate systems are defined on general metric spaces with the purpose of generalizing vector fields on a manifold. Conversion formulae are available between metric and Cartesian coordinates on a Hilbert space. Nagumo's Invariance Theorem is invoked to prove the analogue of the classical Cauchy-Lipschitz Theorem for vector fields on a locally compact coordinatized space. A metric space version of Nagumo's Theorem is one consequence. Examples are given throughout.

math.DS↗

arXiv1996

I-Brane Inflow and Anomalous Couplings on D-Branes

We show that the anomalous couplings of $D$-brane gauge and gravitational fields to Ramond-Ramond tensor potentials can be deduced by a simple anomaly inflow argument applied to intersecting $D$-branes and use this to determine the eight-form gravitational coupling.

hep-th↗