A new framework brings improved text detection across multiple formats and granularities.
― 8 min read
Computer Science - Computer Vision and Pattern Recognition
RSSResearch shows non-humanoid agents can analyze human dance and create movements in sync with music.
― 4 min read
New method improves answers from long videos using innovative techniques.
― 4 min read
A new dataset and detection method tackle the issue of fake video content.
― 6 min read
A new approach to enhance 3D detection accuracy in changing environments.
― 6 min read
New method improves 3D image segmentation consistency across different views.
― 5 min read
A new framework enhances mammogram training for better radiology education.
― 6 min read
New method improves realistic avatar generation using upper and lower body separation.
― 5 min read
A new approach improves model performance with different data sources.
― 6 min read
A new method creates stable 3D objects from images, maintaining appearance and structure.
― 8 min read
This system helps visually impaired individuals shop more independently using a robotic cane.
― 6 min read
Investigating how small errors in training data enhance AI-generated content.
― 5 min read
This article discusses a new simple model for generating audio from images and vice versa.
― 5 min read
This article discusses essential calibration methods for object detectors in critical applications.
― 6 min read
Federated Learning trains models while keeping user data private and secure.
― 5 min read
A novel technique improves 3D modeling accuracy and efficiency.
― 6 min read
Learn how MOFA-Video transforms still images into engaging animations.
― 7 min read
Discover the latest trends in visual data processing and coding.
― 7 min read
A new method for efficiently representing complex 3D images using N-dimensional Gaussians.
― 7 min read
MG-SLAM offers better tracking and mapping for indoor environments using line segments and structured surfaces.
― 7 min read
A study on recognizing actions using few-shot learning and multimodal data.
― 5 min read
Introducing a flexible method for pixel-level anomaly detection in computer vision.
― 6 min read
A new method enables quick and precise recoloring in 3D scenes.
― 6 min read
A new method for running Diffusion Transformers more effectively on smaller devices.
― 6 min read
Discover how SNNs enhance energy efficiency in autonomous driving systems.
― 5 min read
DisDiff aims to prevent misuse of image creation tools while protecting user privacy.
― 4 min read
Research reveals biases in object detection systems affecting safety in autonomous vehicles.
― 5 min read
New techniques improve quality and training for 3D images.
― 6 min read
This study assesses deep learning models in solving complex equations efficiently.
― 6 min read
CLIP exhibits strength in handling data imbalance in visual and language tasks.
― 6 min read
MiDiffusion improves indoor scene creation using floor plans and object attributes.
― 5 min read
MpoxSLDNet offers a promising approach for identifying monkeypox lesions accurately.
― 7 min read
A new approach using neural networks for managing Gaussian scale spaces efficiently.
― 6 min read
Improving sample quality in machine learning through innovative methods.
― 5 min read
A challenge focusing on deep generative models for realistic medical image generation.
― 8 min read
A look at improved techniques for measuring carbon in U.S. forests.
― 6 min read
Using remote sensing to improve forest biomass estimation in China.
― 6 min read
A new approach enhances the ability to describe changes in images accurately.
― 7 min read
A new method improves image generation accuracy from text prompts.
― 7 min read
A new method uses extreme points for effective instance segmentation with minimal annotation.
― 6 min read