Skip to content
View w-xb's full-sized avatar
  • China
  • 10:47 (UTC +08:00)

Block or report w-xb

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
w-xb/README.md

Hi there, I'm Xinbai Wang 👋

Undergraduate @ Harbin Institute of Technology (HIT)

I am an undergraduate student in the 2024 AI Advanced Technology Leader Class at the Harbin Institute of Technology. My research interests are rooted in Computer Vision and Multi-modal AI. I am deeply passionate about Low-Light Image Enhancement (LLIE) and Vision-Language Models (VLMs), and I am actively exploring Vision-Language-Action (VLA) architectures to bridge perception with embodied intelligence.

Email


🏆 Achievements & Publications

  • CVPR 2026 NTIRE Workshop: Co-authored the official technical report for the NTIRE 2026 Efficient Low-Light Image Enhancement (LLIE) Challenge.
  • NTIRE 2026 Challenge Results: Developed MobileIE-6Ch, an ultra-lightweight Retinex-style model with only 101.9K parameters, achieving excellent efficiency-performance trade-offs:
    • Rank 7 in the Main Technical-Report Table
    • Rank 9 in the Full Final-Testing Table
  • Preprint: [Efficient Low-Light Image Enhancement for NTIRE 2026] arXiv

🔭 Current Research & Focus

  • Low-Level Vision (LLIE): Focusing on extreme model compression, lightweight CNNs/ViTs, and real-time image restoration.
  • Vision-Language Models (VLMs): Investigating multi-modal understanding, representation alignment, and generative AI via CLIP, LLaVA, and BLIP.
  • Vision-Language-Action (VLA): Transitioning multi-modal perception into actionable intelligence for next-generation embodied AI systems.

🛠️ Tech Stack & Workflow

  • Frameworks & Deep Learning: PyTorch, torchvision, Transformers, Hugging Face
  • Core Architectures: MobileIE-6Ch, Retinexformer, LLaVA, BLIP
  • AI-Assisted Development: Highly proficient in "Vibecoding"—leveraging AI agents and LLMs to rapidly prototype complex neural network modules, implement experimental pipelines, and accelerate research iterations.

Always open to discussing CV research, MLLMs, Embodied AI (VLA), and potential collaborations! 🚀

Popular repositories Loading

  1. MobileIE-6Ch MobileIE-6Ch Public

    [CVPR 2026 Workshop] Official PyTorch implementation of MobileIE-6Ch for NTIRE 2026 Efficient Low-Light Image Enhancement. 101.9K-parameter Retinex-style model with pretrained checkpoint.

    Python 3

  2. - - Public

    Forked from lu1906429549-bot/-

    让生态修复更科学、更高效。我们的智能系统将前沿技术与生态学知识相结合,为您提供数据驱动的专业恢复方案,旨在加速地球伤痕的愈合,守护绿水青山。

    JavaScript

  3. w-xb w-xb Public