arXiv subjects

John Galeotti

Publications and source records attributed to John Galeotti.

At least 19 recordsLinked to original sources

arXiv2026

Your Model Already Knows Don't Teach It, Learn to Ask It: Soft Prompting for Few-Shot Adaptation of Vision-Language Models

We address few-shot object detection with vision-language models (VLMs) in out-of-domain settings such as aerial, industrial, and medical imagery, using only ten annotated images for supervision. Existing adaptation methods are discrete prompt optimization and LoRA fine-tuning. We revisit a third option: soft prompting, where a small number of continuous prompt tokens are optimized while the pretrained backbone remains frozen. We identify two key design choices. First, placing prompt tokens at the cross-modal boundary between visual and text tokens outperforms other placements (10.0 vs. 8.4 mAP). Second, initializing prompts from the empty space token outperforms semantic and random initialization. With these choices, one to three learned tokens (7,168 parameters on average) match the best LoRA configuration on Roboflow20-VL (14.2 mAP, 10-shot) while training over 20,000x fewer parameters. Soft prompting remains harder to optimize, exhibiting higher variance across random seeds. Unlike LoRA, however, it causes no forgetting: the LoRA rank matching our accuracy reduces NaturalBench VQA accuracy by 35% relative, rising to 56% at the largest rank, whereas soft prompting leaves pretrained performance unchanged. The learned tokens behave like prompts rather than weights. They transfer to a newer model without retraining (+0.8 mAP on Qwen3.5-9B) and can be verbalized into readable prompts competitive with prompt-search methods (matching DetPO and outperforming GEPA). The approach also extends beyond detection. On RoboCasa manipulation tasks, the frozen $\pi_{0.5}$ vision-language-action policy benefits from soft prompting, matching the LoRA baseline on two of three tasks when tokens are placed at the gradient bottleneck. These results suggest modern VLMs already encode much of what is needed for specialized domains; the challenge is learning how to ask.

cs.CV↗

arXiv2026

Attributes Should Come from Images, Not Class Names: Distribution-Conditioned Attribute Selection for Vision-Language Models

A popular route to interpretable zero-shot classification asks a large language model (LLM) to describe each class name and prompts CLIP with the resulting descriptors. We show that these descriptors carry little visual evidence of their own: removing the class name from the prompt collapses ImageNet accuracy from 59.5% to 15.5%. The diagnosis is that the descriptors are conditioned on the label rather than on the images, so they describe the concept in general and mislead exactly when the data shifts; an LLM insists that strawberries are red, but every strawberry in ImageNet-Sketch is a colorless line drawing. We therefore select attributes from the target image collection instead: we score a large attribute pool against the images in CLIP's joint embedding space and keep the top-scoring attributes per class. Selected this way, class-name-free attribute prompts reach 23.8% on ImageNet (against 15.5% for LLM descriptors), the gain holds on four shifted ImageNet variants, and reselecting from the LLM's own pool isolates the selection mechanism as the cause. With one image per class, the selected attributes outperform the prompt-tuning method CoOp by 3 points while fitting in under a minute instead of 14 hours, with no learned soft prompt to obscure the decision. Because the attribute set is chosen by the data, it doubles as a readable summary of a dataset, which we use to describe distribution shift in words. Our code and results are available on our project page: https://ggare-cmu.github.io/AttributeSelect/

cs.CV↗

arXiv2026

Ask Twice, Look Twice: Prompt Echoing Resolves the Question-First Paradox in Vision-Language Models

Where should the question go in a vision-language model (VLM) prompt: before the image or after it? Intuition says before: knowing what is asked should tell the model where to look. Yet across visual question answering benchmarks, question-first prompting consistently underperforms the image-first ordering recommended for frontier VLMs, a phenomenon we term the question-first paradox. We trace this paradox to a conflict between two stages of VLM computation. Logit-lens and attention probes show that question-first prompting steers perception, shifting image patch representations toward question-relevant concepts. But downstream, stranded behind hundreds of image tokens, the question is barely attended by the answer token, which instead commits to image-driven, often wrong answers. Causal attention knockout confirms that the answer reads the question only when it follows the image. This diagnosis yields a training-free fix: question echoing, restating the question on both sides of the image so one copy steers perception while the other is available at answer time. A similar division of labor appears in a fifty-year-old finding on human 'adjunct questions', where repeating a question before and after a passage improves comprehension. Echoing the image as well brings further gains by restoring the whole-image view otherwise lost by a causal decoder. The paradox holds across five open VLMs, costing up to 17.5 group-accuracy points. Echoed prompts recover most of the gap and, on NaturalBench and Winoground, surpass the best single-pass ordering by up to 19 group-accuracy points on Winoground, with no training, fine-tuning, or architecture change. The paradox reveals a tension between steering what a model sees and preserving access to what it was asked; echoing resolves this through prompt design. Project Page: https://rakshanda-cmu.github.io/ask-twice-look-twice/

cs.CV↗

arXiv2026

Do Real-World Datasets Contain Natural Experiments? An Empirical Study Using Causal Feature Selection

In nature, events that affect some individuals or groups but not others constitute an implicit intervention and are known as natural experiments. For example, the COVID-19 pandemic was an intervention by the coronavirus on the sub-population infected with COVID. We ask, do natural experiments occur in existing real-world datasets? If yes, how should we treat them? To detect natural experiments in data, we use causal discovery to recover the underlying causal graph and perform feature selection based on causal links. If downstream performance improves by treating the data as interventional rather than observational, we argue that this suggests the dataset contains natural experiments. We first validate this hypothesis by simulating datasets with and without natural experiments using synthetic graphs. We then perform a systematic empirical evaluation on a large suite of real-world datasets. Our results indicate that real-world datasets do contain natural experiments and we can take advantage of those natural experiments to improve model performance using causal inference. Our work represents the initial foray into this area, offering a preliminary exploration within a limited scope.

cs.AI↗

arXiv2026

Prognostic Value of Lung Ultrasound Biomarkers for Readmission Risk in Congestive Heart Failure: A Pilot Data-Driven Analysis

Hospital readmission within 30 days of discharge is a leading driver of morbidity, mortality, and avoidable healthcare expenditure in congestive heart failure (CHF). Current clinical risk stratification tools rely primarily on non-imaging data and exhibit limited predictive performance. Point-of-care lung ultrasound (LUS) offers a sensitive, noninvasive window into the pulmonary congestion that characterizes CHF decompensation, yet its prognostic utility for readmission prediction remains largely unexplored. We present a pilot feasibility study, the first systematic machine learning study using B-mode LUS acquired during hospitalization to predict 30-day CHF readmission. Quantitative spatiotemporal embeddings are extracted from a pretrained Temporal Shift Module (TSM) ResNet-18 encoder, and interpretable biomarker features are separately evaluated. Through structured ablations over lung view, temporal representation, multi-view fusion, and cross-lung augmentation, we identify the key imaging factors driving readmission risk. Our findings reveal that (1) dependent lower-lung regions (Left-3, Right-3) carry the strongest prognostic signal, consistent with their greater susceptibility to hydrostatic congestion; (2) temporal difference features between sequential examinations substantially outperform single-timepoint representations, highlighting the importance of capturing disease trajectory; and (3) multi-view feature concatenation yields the best overall performance, with our top MLP model achieving an F1 score of 0.80 (95% CI: 0.62-0.96). Biomarker analysis further reveals that pleural-line abnormalities, including breaks and indentations, are as informative as the canonical A-line and B-line markers. These results support POCUS-derived biomarkers as practical, interpretable tools for noninvasive CHF risk stratification.

eess.SP↗

arXiv2026

Meta-Learned Adaptive Optimization for Robust Human Mesh Recovery with Uncertainty-Aware Parameter Updates

Human mesh recovery from single images remains challenging due to inherent depth ambiguity and limited generalization across domains. While recent methods combine regression and optimization approaches, they struggle with poor initialization for test-time refinement and inefficient parameter updates during optimization. We propose a novel meta-learning framework that trains models to produce optimization-friendly initializations while incorporating uncertainty-aware adaptive updates during test-time refinement. Our approach introduces three key innovations: (1) a meta-learning strategy that simulates test-time optimization during training to learn better parameter initializations, (2) a selective parameter caching mechanism that identifies and freezes converged joints to reduce computational overhead, and (3) distribution-based adaptive updates that sample parameter changes from learned distributions, enabling robust exploration while quantifying uncertainty. Additionally, we employ stochastic approximation techniques to handle intractable gradients in complex loss landscapes. Extensive experiments on standard benchmarks demonstrate that our method achieves state-of-the-art performance, reducing MPJPE by 10.3 on 3DPW and 8.0 on Human3.6M compared to strong baselines. Our approach shows superior domain adaptation capabilities with minimal performance degradation across different environmental conditions, while providing meaningful uncertainty estimates that correlate with actual prediction errors. Combining meta-learning and adaptive optimization enables accurate mesh recovery and robust generalization to challenging scenarios.

cs.CV↗

arXiv2026

DetPO: In-Context Learning with Multi-Modal LLMs for Few-Shot Object Detection

Multi-Modal LLMs (MLLMs) demonstrate strong visual grounding capabilities on popular object detection benchmarks like OdinW-13 and RefCOCO. However, state-of-the-art models still struggle to generalize to out-of-distribution classes, tasks and imaging modalities not typically found in their pre-training. While in-context prompting is a common strategy to improve performance across diverse tasks, we find that it often yields lower detection accuracy than prompting with class names alone. This suggests that current MLLMs cannot yet effectively leverage few-shot visual examples and rich textual descriptions for object detection. Since frontier MLLMs are typically only accessible via APIs, and state-of-the-art open-weights models are prohibitively expensive to fine-tune on consumer-grade hardware, we instead explore black-box prompt optimization for few-shot object detection. To this end, we propose Detection Prompt Optimization (DetPO), a gradient-free test-time optimization approach that refines text-only prompts by maximizing detection accuracy on few-shot visual training examples while calibrating prediction confidence. Our proposed approach yields consistent improvements across generalist MLLMs on Roboflow20-VL and LVIS, outperforming prior black-box approaches by up to 9.7 mAP. Our code and optimized prompts are available at https://ggare-cmu.github.io/DetPO/

cs.CV↗

arXiv2025

Label Uncertainty for Ultrasound Segmentation

In medical imaging, inter-observer variability among radiologists often introduces label uncertainty, particularly in modalities where visual interpretation is subjective. Lung ultrasound (LUS) is a prime example-it frequently presents a mixture of highly ambiguous regions and clearly discernible structures, making consistent annotation challenging even for experienced clinicians. In this work, we introduce a novel approach to both labeling and training AI models using expert-supplied, per-pixel confidence values. Rather than treating annotations as absolute ground truth, we design a data annotation protocol that captures the confidence that radiologists have in each labeled region, modeling the inherent aleatoric uncertainty present in real-world clinical data. We demonstrate that incorporating these confidence values during training leads to improved segmentation performance. More importantly, we show that this enhanced segmentation quality translates into better performance on downstream clinically-critical tasks-specifically, estimating S/F oxygenation ratio values, classifying S/F ratio change, and predicting 30-day patient readmission. While we empirically evaluate many methods for exposing the uncertainty to the learning model, we find that a simple approach that trains a model on binarized labels obtained with a (60%) confidence threshold works well. Importantly, high thresholds work far better than a naive approach of a 50% threshold, indicating that training on very confident pixels is far more effective. Our study systematically investigates the impact of training with varying confidence thresholds, comparing not only segmentation metrics but also downstream clinical outcomes. These results suggest that label confidence is a valuable signal that, when properly leveraged, can significantly enhance the reliability and clinical utility of AI in medical imaging.

eess.IV↗

arXiv2025

Automatic Cannulation of Femoral Vessels in a Porcine Shock Model

Rapid and reliable vascular access is critical in trauma and critical care. Central vascular catheterization enables high-volume resuscitation, hemodynamic monitoring, and advanced interventions like ECMO and REBOA. While peripheral access is common, central access is often necessary but requires specialized ultrasound-guided skills, posing challenges in prehospital settings. The complexity arises from deep target vessels and the precision needed for needle placement. Traditional techniques, like the Seldinger method, demand expertise to avoid complications. Despite its importance, ultrasound-guided central access is underutilized due to limited field expertise. While autonomous needle insertion has been explored for peripheral vessels, only semi-autonomous methods exist for femoral access. This work advances toward full automation, integrating robotic ultrasound for minimally invasive emergency procedures. Our key contribution is the successful femoral vein and artery cannulation in a porcine hemorrhagic shock model.

cs.RO↗

arXiv2024

LEARNER: Contrastive Pretraining for Learning Fine-Grained Patient Progression from Coarse Inter-Patient Labels

Predicting whether a treatment leads to meaningful improvement is a central challenge in personalized medicine, particularly when disease progression manifests as subtle visual changes over time. While data-driven deep learning (DL) offers a promising route to automate such predictions, acquiring large-scale longitudinal data for each individual patient remains impractical. To address this limitation, we explore whether inter-patient variability can serve as a proxy for learning intra-patient progression. We propose LEARNER, a contrastive pretraining framework that leverages coarsely labeled inter-patient data to learn fine-grained, patient-specific representations. Using lung ultrasound (LUS) and brain MRI datasets, we demonstrate that contrastive objectives trained on coarse inter-patient differences enable models to capture subtle intra-patient changes associated with treatment response. Across both modalities, our approach improves downstream classification accuracy and F1-score compared to standard MSE pretraining, highlighting the potential of inter-patient contrastive learning for individualized outcome prediction.

eess.IV↗

arXiv2023

Motion Informed Needle Segmentation in Ultrasound Images

Segmenting a moving needle in ultrasound images is challenging due to the presence of artifacts, noise, and needle occlusion. This task becomes even more demanding in scenarios where data availability is limited. In this paper, we present a novel approach for needle segmentation for 2D ultrasound that combines classical Kalman Filter (KF) techniques with data-driven learning, incorporating both needle features and needle motion. Our method offers three key contributions. First, we propose a compatible framework that seamlessly integrates into commonly used encoder-decoder style architectures. Second, we demonstrate superior performance compared to recent state-of-the-art needle segmentation models using our novel convolutional neural network (CNN) based KF-inspired block, achieving a 15\% reduction in pixel-wise needle tip error and an 8\% reduction in length error. Third, to our knowledge we are the first to implement a learnable filter to incorporate non-linear needle motion for improving needle segmentation.

eess.IV↗

arXiv2023

Unsupervised Deformable Ultrasound Image Registration and Its Application for Vessel Segmentation

This paper presents a deep-learning model for deformable registration of ultrasound images at online rates, which we call U-RAFT. As its name suggests, U-RAFT is based on RAFT, a convolutional neural network for estimating optical flow. U-RAFT, however, can be trained in an unsupervised manner and can generate synthetic images for training vessel segmentation models. We propose and compare the registration quality of different loss functions for training U-RAFT. We also show how our approach, together with a robot performing force-controlled scans, can be used to generate synthetic deformed images to significantly expand the size of a femoral vessel segmentation training dataset without the need for additional manual labeling. We validate our approach on both a silicone human tissue phantom as well as on in-vivo porcine images. We show that U-RAFT generates synthetic ultrasound images with 98% and 81% structural similarity index measure (SSIM) to the real ultrasound images for the phantom and porcine datasets, respectively. We also demonstrate that synthetic deformed images from U-RAFT can be used as a data augmentation technique for vessel segmentation models to improve intersection-over-union (IoU) segmentation performance

eess.IV↗

arXiv2023

Unsupervised Deformable Image Registration for Respiratory Motion Compensation in Ultrasound Images

In this paper, we present a novel deep-learning model for deformable registration of ultrasound images and an unsupervised approach to training this model. Our network employs recurrent all-pairs field transforms (RAFT) and a spatial transformer network (STN) to generate displacement fields at online rates (apprx. 30 Hz) and accurately track pixel movement. We call our approach unsupervised recurrent all-pairs field transforms (U-RAFT). In this work, we use U-RAFT to track pixels in a sequence of ultrasound images to cancel out respiratory motion in lung ultrasound images. We demonstrate our method on in-vivo porcine lung videos. We show a reduction of 76% in average pixel movement in the porcine dataset using respiratory motion compensation strategy. We believe U-RAFT is a promising tool for compensating different kinds of motions like respiration and heartbeat in ultrasound images of deformable tissue.

eess.IV↗

arXiv2021

Toward Robotically Automated Femoral Vascular Access

Advanced resuscitative technologies, such as Extra Corporeal Membrane Oxygenation (ECMO) cannulation or Resuscitative Endovascular Balloon Occlusion of the Aorta (REBOA), are technically difficult even for skilled medical personnel. This paper describes the core technologies that comprise a teleoperated system capable of granting femoral vascular access, which is an important step in both of these procedures and a major roadblock in their wider use in the field. These technologies include a kinematic manipulator, various sensing modalities, and a user interface. In addition, we evaluate our system on a surgical phantom as well as in-vivo porcine experiments. These resulted in, to the best of our knowledge, the first robot-assisted arterial catheterizations; a major step towards our eventual goal of automatic catheter insertion through the Seldinger technique.

cs.RO↗

arXiv2020

Good and Bad Boundaries in Ultrasound Compounding: Preserving Anatomic Boundaries While Suppressing Artifacts

Ultrasound 3D compounding is important for volumetric reconstruction, but as of yet there is no consensus on best practices for compounding. Ultrasound images depend on probe direction and the path sound waves pass through, so when multiple intersecting B-scans of the same spot from different perspectives yield different pixel values, there is not a single, ideal representation for compounding (i.e. combining) the overlapping pixel values. Current popular methods inevitably suppress or altogether leave out bright or dark regions that are useful, and potentially introduce new artifacts. In this work, we establish a new algorithm to compound the overlapping pixels from different view points in ultrasound. We uniquely leverage Laplacian and Gaussian Pyramids to preserve the maximum boundary contrast without overemphasizing noise and speckle. We evaluate our algorithm by comparing ours with previous algorithms, and we show that our approach not only preserves both light and dark details, but also somewhat suppresses artifacts, rather than amplifying them.

eess.IV↗

arXiv2020

Ultrasound Confidence Maps of Intensity and Structure Based on Directed Acyclic Graphs and Artifact Models

Ultrasound imaging has been improving, but continues to suffer from inherent artifacts that are challenging to model, such as attenuation, shadowing, diffraction, speckle, etc. These artifacts can potentially confuse image analysis algorithms unless an attempt is made to assess the certainty of individual pixel values. Our novel confidence algorithms analyze pixel values using a directed acyclic graph based on acoustic physical properties of ultrasound imaging. We demonstrate unique capabilities of our approach and compare it against previous confidence-measurement algorithms for shadow-detection and image-compounding tasks.

eess.IV↗

arXiv2020

Weakly- and Semi-Supervised Probabilistic Segmentation and Quantification of Ultrasound Needle-Reverberation Artifacts to Allow Better AI Understanding of Tissue Beneath Needles

Ultrasound image quality has continually been improving. However, when needles or other metallic objects are operating inside the tissue, the resulting reverberation artifacts can severely corrupt the surrounding image quality. Such effects are challenging for existing computer vision algorithms for medical image analysis. Needle reverberation artifacts can be hard to identify at times and affect various pixel values to different degrees. The boundaries of such artifacts are ambiguous, leading to disagreement among human experts labeling the artifacts. We propose a weakly- and semi-supervised, probabilistic needle-and-reverberation-artifact segmentation algorithm to separate the desired tissue-based pixel values from the superimposed artifacts. Our method models the intensity decay of artifact intensities and is designed to minimize the human labeling error. We demonstrate the applicability of the approach and compare it against other segmentation algorithms. Our method is capable of differentiating between the reverberations from artifact-free patches as well as of modeling the intensity fall-off in the artifacts. Our method matches state-of-the-art artifact segmentation performance and sets a new standard in estimating the per-pixel contributions of artifact vs underlying anatomy, especially in the immediately adjacent regions between reverberation lines. Our algorithm is also able to improve the performance downstream image analysis algorithms.

eess.IV↗

arXiv2020

A Study of Domain Generalization on Ultrasound-based Multi-Class Segmentation of Arteries, Veins, Ligaments, and Nerves Using Transfer Learning

Identifying landmarks in the femoral area is crucial for ultrasound (US) -based robot-guided catheter insertion, and their presentation varies when imaged with different scanners. As such, the performance of past deep learning-based approaches is also narrowly limited to the training data distribution; this can be circumvented by fine-tuning all or part of the model, yet the effects of fine-tuning are seldom discussed. In this work, we study the US-based segmentation of multiple classes through transfer learning by fine-tuning different contiguous blocks within the model, and evaluating on a gamut of US data from different scanners and settings. We propose a simple method for predicting generalization on unseen datasets and observe statistically significant differences between the fine-tuning methods while working towards domain generalization.

cs.CV↗