Chenglong Li

SCDT outperforms existing methods by unifying feature learning for both modality-missing and complete scenarios without altering architecture or parameters.

A groundbreaking benchmark that captures the complexities of real-world dynamic tracking with 795K RGBT frame pairs and extensive annotations for robust evaluation.

Robust UAV tracking is revolutionized with SARLA, which adeptly manages sudden appearance and spatial shifts caused by modality switching.

EV-MoE not only enhances feature representation but also introduces a large-scale benchmark that redefines multi-query vehicle ReID evaluation in complex environments.

Achieving state-of-the-art accuracy in RGB-Thermal video object detection, DHNet tackles spatial misalignment with innovative dual-correlation learning.

Transformers secretly act like the power method, concentrating token embeddings along a dominant eigenvector with each layer.

Existing referring detection models fall apart when confronted with the scale variations and complexity of aerial imagery, but a new framework closes the gap.

Can AI transform a grumpy cat meme into a beacon of positivity while keeping the cat recognizable?

Forget retraining your agent: Steve-Evolving distills execution failures into executable guardrails and successes into reusable skills, injecting them into an LLM planner for continual, parameter-free improvement.