Tag: computer vision
Browse our exclusive articles!
Hierarchical Pre-Training of Vision Encoders with LLMs
Discover HIVE, a novel hierarchical pre-training method that enhances vision encoders with large language models for superior multimodal AI performance.
FERA: Pose-Based Multimedia Decision Support for Foil Fencing
Discover FERA, a pose-based framework enhancing rule-grounded multimedia decision support with a foil fencing case study for accurate, explainable decision...
MELT: Enhancing Composed Image Retrieval with Frequency-Rarity Balance
Discover MELT, a novel network improving composed image retrieval by balancing modification frequency and rarity for better accuracy and denoising.
WorldFlow3D: Advanced Unbounded 3D World Generation
Discover WorldFlow3D, a novel flow matching method for high-quality, controllable unbounded 3D world generation in vision and robotics.
Zero-Shot Vision-Language Reranking Boosts Geolocalization
Improve cross-view geolocalization accuracy using zero-shot vision-language reranking with pairwise comparison strategies.