Possible advisorsa guess from early papers, not confirmed

• Guosheng Lin

  Possible advisor · last author on 4 of their early first-author papers, 2023–2024

  Suggested from co-authorship

Is this you? Claim this profile to confirm or dismiss it.

• Self-Training Vision Language BERTs With a Unified Conditional Model

  Xiaofeng Yang, Fengmao Lv, Fayao Liu, Guosheng Lin

  IEEE Transactions on Circuits and Systems for Video Technology · 2023

  A self-training approach that allows training VL-BERTs from unlabeled image data and is able to get competitive or even better performances compared to the models of similar model size trained with 3 million extra image data.

  22
• Effective End-to-End Vision Language Pretraining With Semantic Visual Loss

  Xiaofeng Yang, Fayao Liu, Guosheng Lin

  IEEE Transactions on Multimedia · 2023

  This paper systematically study how to leverage auxiliary visual pretraining tasks to help training end-to-end vision language models and introduces three types of visual losses that enable much faster convergence and better finetuning accuracy.

  18
• Neural Logic Vision Language Explainer

  Xiaofeng Yang, Fayao Liu, Guosheng Lin

  IEEE Transactions on Multimedia · 2023

  7
• Detail++: Training-Free Detail Enhancer for T2I Diffusion Models

  Lifeng Chen, Jin’er Wang, Zihao Pan, Beier Zhu, Xiaofeng Yang, Chi Zhang

  IEEE Transactions on Image Processing · 2026

  This work decomposes a complex prompt into a sequence of simplified sub-prompts, guiding the generation process in stages, and introduces a novel Progressive Detail Injection (PDI) strategy to address this limitation.

  6
• Low-latency and energy-efficient FPGA accelerator for sparse neural networks in edge LiDAR-based 3D object detection

  Jinyi Li, Yuchen Chen, Xiaofeng Yang, Hengrui Hu, Shengli Lu

  The Journal of Supercomputing · 2025

  A low-latency FPGA accelerator optimized for sparse neural networks using adaptive hybrid sparse convolution (AHSC), which uses a lightweight predictor to dynamically choose between feature pruning, submanifold sparse convolution, or standard convolution, balancing accuracy and sparsity is presented.

  3
• Neural Radiance Selector: Find the best 2D representations of 3D data for CLIP based 3D tasks

  Xiaofeng Yang, Fayao Liu, Guosheng Lin

  Knowledge-Based Systems · 2024

  3
• Learning language to symbol and language to vision mapping for visual grounding

  Su He, Xiaofeng Yang, Guosheng Lin

  Image and Vision Computing · 2022

  3
• Cephalometric radiograph generation from 3D dental CBCT images with automatic positioning

  Dan Xia, Zhenhua Yang, Hongwei Cao, Xiaofeng Yang, Hongyan Qu, Fangjun Tian, Xiaochuan Pan

  SPIE Medical Imaging: Physics of Medical Imaging · 2025

  A new method for generating digital cephalometric radiographs from 3D images acquired with a dental CBCT system is presented and the preliminary results reveal that the proposed method can yield cephalometric radiographs with improved contrast and landmarks of clinical relevance.

  –