Latest Articles for Computer Vision

Discover EVE, a model improving understanding of images and text.

FG-Net improves automatic detection of facial emotions using efficient techniques.

Strategies to improve image classification by minimizing background influence.

RefEgo dataset enhances video object recognition through natural language instructions.

A new method improves efficiency and accuracy in visual localization tasks.

A model predicts motion of rigid bodies using images, addressing mass distribution challenges.

APLA improves video generation by ensuring frame consistency and detail retention.

AccFlow uses backward accumulation to improve long-range optical flow estimation.

A look at how TAL models work with limited data and computing power.

Learn how data factors influence CNN efficiency for image tasks.

MapPrior improves BEV perception, enhancing accuracy and safety for autonomous vehicles.

New technique improves 3D detection accuracy using a single camera.

A new method combines depth estimation and segmentation to enhance autonomous vehicle safety.

Research focuses on improving models that connect visuals and text through language understanding.

FaceTouch tracks hand-to-face contacts to help reduce disease spread.

New markers improve shape tracking on smooth surfaces.

CS-Mixer offers a new way to process images by combining information from different scales.

A new method improves landmark detection by masking distractions in images.

JointFormer enhances VOS by integrating feature extraction, matching, and memory management.

A new self-supervised method enhances image resolution without paired data.

RestNet improves segmentation tasks with limited data across different domains.

A new method enhances Transformer models for better medical image analysis.

New model improves recognition of facial expressions in videos.

SaEnet enhances CNN performance by focusing on essential features in images.

A new approach improves bundle adjustment speed using dynamic damping.

OVDEval benchmark challenges OVD models to improve their evaluation methods.

A method for improving image segmentation using contextual information from pixels.

SOGDet improves object detection by considering environmental context for autonomous driving.

A new method enhances localization accuracy using line data from panoramic images.

This study examines how eye tracking enhances the performance of Vision Transformers in driving tasks.

Researchers develop weight masks to help models retain knowledge while learning new tasks.

A new method improves pose estimation using only target domain data.

This article explores the role of Transformers in image restoration and their vulnerabilities to adversarial attacks.

A new method enhances 3D face modeling using everyday images.

A look into the methods and applications of Human Pose Estimation.

A new method enhances OOD detection using normalizing flows and manifold learning.

DiffI2I enhances image-to-image translation with improved accuracy and efficiency.

A new method to enhance depth mapping in robotics and AR.

A new tool estimates how neural networks react to input changes, crucial for safety.

Research on improving model performance with varied point cloud datasets.