Mengqi Lei (雷孟奇)

My research interests lie in visual intelligence and multimodal learning. I work on vision-language models, vision foundation models, object detection, object counting, high-order representation learning, etc. My objective is to develop a more general, comprehensive, and efficient framework for visual and multimodal representation learning.

Showing 5 selected publications.

Research Papers

Count Anything

Mengqi Lei, Shuokun Cheng, Wei Bao, Shaoyi Du, Jun-Hai Yong, Siqi Li, and Yue Gao

arXiv, 2026

• Visual Intelligence
• Multimodal Learning

A generalist text-guided object-counting model that combines region-level sparse and pixel-level dense point prediction to localize and count targets across diverse visual domains.

project / paper / code / model / demo / bibtex

Hypergraph as Language

Mengqi Lei, Guohuan Xie, Shihui Ying, Shaoyi Du, Jun-Hai Yong, Siqi Li, and Yue Gao

arXiv, 2026

• Multimodal Learning

A hypergraph-native alignment framework that makes query-centered high-order association structures directly consumable by large language models.

paper / code / bibtex

YOLOv13: Real-Time Object Detection with Hypergraph-Enhanced Adaptive Visual Perception

Mengqi Lei, Siqi Li, Yihong Wu, Han Hu, You Zhou, Xinhu Zheng, Guiguang Ding, Shaoyi Du, Zongze Wu, and Yue Gao

arXiv, 2025

• Visual Intelligence

A real-time detector that uses hypergraph-based adaptive correlation enhancement and full-pipeline feature aggregation.

paper / code / bibtex

SoftHGNN: Soft Hypergraph Neural Networks for General Visual Recognition

Mengqi Lei, Yihong Wu, Siqi Li, Xinhu Zheng, Juan Wang, Shaoyi Du, and Yue Gao

IJCV, 2026

• Visual Intelligence

Soft, learnable hyperedges model continuous high-order visual relationships across classification, crowd counting, and object detection.

publisher / arXiv / code / bibtex

ConDSeg: A General Medical Image Segmentation Framework via Contrast-Driven Feature Enhancement

Mengqi Lei, Haochen Wu, Xinhua Lv, and Xin Wang

AAAI, 2025

• Visual Intelligence

A contrast-driven feature enhancement framework for general medical image segmentation across diverse imaging conditions.

paper / code / bibtex

No publications match this category.