Latest Research in Computer Vision and Pattern Recognition
78 research papers · 2026 median publication year
Top Research Topics in Computer Vision and Pattern Recognition
- Computer Vision and Pattern Recognition — 64 papers
- Multimodal Machine Learning Applications — 5 papers
- Computation and Language — 2 papers
- Multimedia — 2 papers
- Information Retrieval — 1 papers
- Intelligence, Security, War Strategy — 1 papers
- Machine Learning — 1 papers
- Artificial Intelligence — 1 papers
- Audio and Speech Processing — 1 papers
Highest-Cited Papers
- CLIP-CC-Bench: Evaluating Paragraph-Level Video Descriptions in Video-Language Models
- Persistent Object Narratives for Token-Efficient Video Language Models
- GROVE: Growing and Reasoning over Temporally Stratified Memory from Streaming Video Experience
- Predict, Then Retrieve: Cross-Instance Future-State Retrieval from Video Prefixes
- ZoomV: Temporal Zoom-in for Efficient Long Video Understanding
- Reading Between the Frames: Interpreting Implicit and Non-literal Meaning in Social Media Videos
- REZE: Recognition-Based Zero-Shot Extraction for Video Temporal Grounding
- GSTEP: Global Spatio-Temporal Density-Driven Visual Token Pruning for Efficient Video Large Language Models
- Adaptive Modality Reliability Diagnosis and Restoration for Robust Multimodal Intent Recognition
- When and Where to Look: Adaptive Visual Evidence Scheduling for Efficient Long Video Understanding
- TBSG-Net: Temporal Bipartite Scene Graph Network for Fine-Grained Video Moment Retrieval
- EgoIntent: A Pre-Outcome Micro-Step Benchmark for Understanding What, Why, and Next
- WaveZip: Wavelet-Driven Space-Time Decoupling for Video Token Condensation
- Ground, Cover, and Refine: Evidence-Centric Frame Selection for Long-Video Question Answering
- CAVE: Competence-Aware Visual Boundary Evidence Alignment for Video Temporal Grounding
- NarrativeTrack: Evaluating Entity-Centric Reasoning for Narrative Understanding
- Rethinking Video Token Compression with a Global Codebook: Learning Once, Compressing Everywhere
- ObjectStream: Latent Objects as Memory Anchors for Streaming Video Understanding
- PHA-Net: Prototype-based Hierarchical Alignment Network for Text-Video Retrieval
- Coverage-Driven Adaptive Keyframe Selection for Video Understanding