Search for: All records

Amodal segmentation is an image-based algorithm that aims to predict masks for both visible and occluded parts of objects. Existing methods typically rely on supervised learning with annotated amodal masks or synthetic data. The effectiveness of these methods relies heavily on the quality of the datasets. This dependence can unintentionally restrict their generalization capabilities due to insufficient diversity and size. Although existing zero-shot methods perform well on their reported datasets, their performance does not necessarily transfer to other datasets. We propose a tuning-free approach that re-purposes diffusion-based inpainting foundation models for amodal segmentation. Our approach is motivated by the “occlusion-free bias” of inpainting models, i.e., the inpainted objects tend to be complete and without occlusions. We reconstruct the occluded regions of an object via inpainting and then apply segmentation, all without additional training or fine-tuning. Experiments on five datasets, three previously unreported, demonstrate the generalizability of our approach. On average, our approach achieves 5.3% more accurate masks in mIoU compared to the publicly available state-of-the-art, pix2gestalt. 

Finding where the trees are in a city and monitoring any changes are essential for sustainable urban management. Historically, urban forests are mainly inventoried via manual processes often limited to public lands. Leveraging advances in computing, we present a novel generative artificial intelligence (AI) method along with a first-ever national-scale dataset, to automatically localize trees in cities across the nation using satellite imagery. Our monitoring approach is fully automated and can be completed for 330 U.S. cities within less than a day of computing, enabling actionable knowledge of changes in urban trees and supporting sustainable development decisions. We successfully localized and counted over 278 million trees, achieving an average tree count accuracy of 92.5% and spatial accuracy of 1.5m for 2024–2025. Our computational approach allows for novel nationwide analysis to be performed. For example, we can localize approximately 117 million trees on private lands and 161 million on public lands. Further, we show and quantify that urban tree distribution exhibits strong spatial disparity, with low-income communities having substantially fewer trees and less canopy cover than others. In addition, we compare tree count and layouts before and after multiple major events (e.g., major fires and destructive weather phenomena). Overall, our approach enhances computational urban planning, including weather and extreme event forecasting, for the development of sustainable cities. 

Model immunization is an emerging direction that aims to mitigate the potential risk of misuse associated with open-sourced models and advancing adaptation methods. The idea is to make the released models' weights difficult to fine-tune on certain harmful applications, hence the name immunized. Recent work on model immunization focuses on the single-concept setting. However, in real-world situations, models need to be immunized against multiple concepts. To address this gap, we propose an immunization algorithm that, simultaneously, learns a single difficult initialization for adaptation methods over a set of concepts. We achieve this by incorporating a differentiable merging layer that combines a set of model weights adapted over multiple concepts.In our experiments, we demonstrate the effectiveness of multi-concept immunization by generalizing prior work's experiment setup of re-learning and personalization adaptation to multiple concepts. 

Abstract Atmospheric aerosol particles that contain water‐soluble components can absorb water vapor in humid environments and form either haze particles or cloud droplets, depending on supersaturation conditions. Laboratory and in situ measurements have shown that haze particles and cloud droplets often coexist and compete for available water vapor in shallow clouds and fogs, especially under polluted conditions. It is expected that more aerosol particles can form more haze particles and cloud droplets, so that the haze and cloud number concentrations are positively correlated. However, recent large‐eddy simulations show that haze and cloud number concentration can be negatively correlated under extremely polluted conditions. Haze–cloud interactions across different environmental settings remain poorly understood. In this study, experiments in a convection cloud chamber with the same aerosol injection rate show that, as supersaturation forcing increases (i.e., changing from polluted to clean conditions), the covariance between haze and cloud droplet number concentration changes from negative to positive and finally to zero. Large‐eddy simulations (LES) of the cloud chamber with a fixed supersaturation forcing but varying aerosol injection rates show a similar result for the haze–cloud correlation: near zero in clean, mean‐supersaturation‐dominated activation conditions; positive in moderate, supersaturation‐fluctuation‐influenced activation conditions; and negative in polluted, supersaturation‐fluctuation‐dominated activation conditions. A theoretical covariance framework was developed to interpret this behavior based on the relative magnitude and signs of the correlations between supersaturation and the populations of haze and cloud droplets. Significantly, experiments, LES, and theory all yield the same three‐regime behavior for the sign of the haze–cloud covariance. Our results show that the haze–cloud covariance remains robust and easily measurable, thereby providing a useful metric for regime identification in the atmosphere, improving regime‐aware parameterizations, and informing aerosol interventions such as fog dispersion, rainfall enhancement, and albedo modification. 

Abstract. This study delves into the small-scale temperature structure inside the turbulent convection Π Chamber under three temperature differences (10, 15, and 20 K) at Rayleigh number Ra∼109 and Prandtl number Pr≈0.7. We performed high-frequency measurements (2 kHz) with the UltraFast Thermometer (UFT) at selected points along the vertical axis. The miniaturized design of the sensor with a resistive platinum-coated tungsten wire, 2.5 µm thick and 3 mm long, mounted on a miniature wire probe, allowed for vertically undisturbed temperature profiling through the chamber's depth spanning from 8 cm above the bottom to 5 cm below the top. The collected data, consisting of 19 and 3 min time series, were used to investigate the variability of the temperature field within the chamber, aiming to better address scientific questions related to its primary objective: understanding small-scale aerosol–cloud interactions. The analyses reveal substantial variability in both variance and skewness of temperature distributions near the top and bottom plates and in the bulk (central) region, which were linked to local thermal plume dynamics. We also identified three spectral regimes termed “inertial range” (slopes of ∼-7/5), “transition range” (slopes of ∼-3), and “dissipative range”, characterized by slopes of ∼-7. Furthermore, the analysis showed a power law relationship between the periodicity of large-scale circulation (LSC) and the temperature difference. Notably, the experimental results are in good agreement with direct numerical simulation (DNS) conducted under similar thermodynamic conditions, illustrating a comparative analysis of this nature. 

Abstract The task of crafting procedural programs capable of generating structurally valid 3D shapes easily and intuitively remains an elusive goal in computer vision and graphics. Within the graphics community, generating procedural 3D models has shifted to using node graph systems. They allow the artist to create complex shapes and animations through visual programming. Being a high‐level design tool, they made procedural 3D modelling more accessible. However, crafting those node graphs demands expertise and training. We present GeoCode, a novel framework designed to extend an existing node graph system and significantly lower the bar for the creation of new procedural 3D shape programs. Our approach meticulously balances expressiveness and generalization for part‐based shapes. We propose a curated set of new geometric building blocks that are expressive and reusable across domains. We showcase three innovative and expressive programs developed through our technique and geometric building blocks. Our programs enforce intricate rules, empowering users to execute intuitive high‐level parameter edits that seamlessly propagate throughout the entire shape at a lower level while maintaining its validity. To evaluate the user‐friendliness of our geometric building blocks among non‐experts, we conduct a user study that demonstrates their ease of use and highlights their applicability across diverse domains. Empirical evidence shows the superior accuracy of GeoCode in inferring and recovering 3D shapes compared to an existing competitor. Furthermore, our method demonstrates superior expressiveness compared to alternatives that utilize coarse primitives. Notably, we illustrate the ability to execute controllable local and global shape manipulations. Our code, programs, datasets and Blender add‐on are available athttps://github.com/threedle/GeoCode.