Computer Vision

Object-centric occupancy completion boosts 3D object detection accuracy by using temporal information from long sequences to precisely reconstruct object shapes, particularly for incomplete or distant…

This paper introduces novel data augmentation modules that dynamically adjust the frequency characteristics of datasets, preventing neural networks from over-relying on simple frequency-based shortcut…

SGNet cultivates general segmentation models for ultra images by integrating surrounding context, achieving significant performance improvements across various datasets.

Dynamic NLOS imaging gets a speed boost! New ST-Mamba method leverages temporal consistency across frames for high-resolution video reconstruction, overcoming speed limitations of traditional methods.

Boosting 3D human pose estimation: Biomechanical Pose Generator and Binary Depth Coordinates enhance accuracy and scalability.

TopoFR enhances face recognition by aligning topological structures between input and latent spaces. Using persistent homology, it preserves crucial data structure info, overcoming overfitting. A har…

This paper introduces novel metrics for visual emotion recognition evaluation, considering the psychological distance between emotions to better reflect human perception, improving the assessment of m…

TinyLUT achieves 10x lower memory consumption and superior accuracy in image restoration on edge devices using innovative separable mapping and dynamic discretization of LUTs.

Terra, a novel time-varying low-rank adapter, enables effective cross-domain fine-tuning of diffusion models by creating a continuous parameter manifold, facilitating efficient knowledge sharing and g…

Using real images retrieved from a generator’s training data outperforms using synthetic images generated by that same model for image classification.

R3GAN, a minimalist GAN baseline, surpasses state-of-the-art models by using a novel regularized relativistic GAN loss and modern architectures, proving GANs can be trained efficiently without relying…

TFS-NeRF: A template-free neural radiance field efficiently reconstructs semantically separable 3D geometries of dynamic scenes featuring multiple interacting entities from sparse RGB videos.

Test-Time Dynamic Image Fusion (TTD) paradigm provably improves image fusion by dynamically weighting source data based on their relative dominance, reducing generalization error without extra trainin…

Low-rank block trifocal tensor unlocks accurate, efficient camera pose synchronization.

ConVRT: A novel framework restores turbulence-distorted videos by decoupling spatial and temporal information in a neural representation, achieving temporally consistent mitigation.

This research introduces a template-free articulated Gaussian splatting method for real-time dynamic view synthesis, automatically discovering object skeletons from videos to enable reposing.

TARSS-Net: A novel temporal-aware radar semantic segmentation network uses a data-driven approach to aggregate temporal information, enhancing accuracy and performance.

TARP-VP reveals a surprising lack of trade-off between adversarial robustness and privacy for label mapping visual prompting models, showing that transferred adversarial training significantly improve…

APCT: a novel architecture enhances 3D point cloud recognition by using an adversarial feature erasing mechanism to improve global structure capture and robustness against real-world corruptions.

TAPTRv2 enhances point tracking by introducing an attention-based position update, eliminating cost-volume reliance for improved accuracy and efficiency.