LSKNet: A foundation lightweight backbone for remote sensing (IJCV)

Yuxuan Li, Xiang Li#, Yimian Dai, Qibin Hou, Li Liu, Yongxiang Liu, Ming-Ming Cheng, Jian Yang#

[Paper] [BibTex] [Demo Video] [Code]

LSKNet can dynamically adjust its large spatial receptive field to better model the ranging context of various categories of objects in remote sensing scenarios.

APF-GAN:Exploring asymmetric pre-training and fine-tuning strategy for conditional generativeadversarial network (CVMJ)

Yuxuan Li, Lingfeng Yang, Xiang Li#

[Paper] [Code]

The APF-GAN model improves GAN-based image segmentation, surpassing GauGAN and winning the Second Jittor AI Challenge.

Large Selective Kernel Network for Remote Sensing Object Detection (ICCV 2023)

Yuxuan Li, Qibin Hou, Zhaohui Zheng, Ming-Ming Cheng, Jian Yang#, Xiang Li#

[Paper] [BibTex] [Demo Video] [Report/Forum] [Code] [知乎专栏]

LSKNet can dynamically adjust its large spatial receptive field to better model the ranging context of various categories of objects in remote sensing scenarios.

SARDet-100K: Towards Open-Source Benchmark and ToolKit for Large-Scale SAR Object Detection (NeurIPS 2024 Spotlight)

Yuxuan Li, Xiang Li#, Weijie Li, Qibin Hou, Li Liu, Ming-Ming Cheng, Jian Yang#

[Paper] [BibTex] [Code] [知乎专栏]

SARDet-100K, the first large-scale multi-class SAR object detection dataset, along with the proposed MSFA pretraining framework, addresses domain and model disparities, significantly improving SAR object detection performance and advancing the field.

Spatial group-wise enhance: Enhancing semantic feature learning in CNN (ACCV 2022)

Yuxuan Li, Xiang Li#, Jian Yang

[Paper] [BibTex] [[Code]] [知乎专栏]

The Spatial Group-wise Enhance (SGE) module improves CNN performance by generating accurate spatial attention masks through local-global similarity within semantic groups, achieving notable accuracy gains on ImageNet and COCO tasks with minimal overhead.

• 2022, Champion of Jittor AI competition, 50,000 RMB bonus (1st from 154 teams).

  • summaries
  • solutions
  • code
  • certificate

• 2022, First Prize of the 5th Creative Open-Source competition, 50,000 RMB bonus.

  • summaries
  • solutions
  • code
  • certificate

• 2022, Second place of IACC International Algorithm Case Competition, 100,000 RMB bonus (2nd from 116 teams).

  • summaries
  • solutions
  • code
  • certificate

• 2023, Third prize of Jittor AI competition, 10,000 RMB bonus (4th from 109 teams).

  • solutions
  • code
  • certificate

• 2023, Third prize of “Jilin-1” Satellite Remote Sensing Application Youth Innovation and Entrepreneurship Competition, (5th from 110 teams).

  • code
  • summaries

• 2023, Outstanding Presentation Award The 3rd Academic Forum for Graduate Students of Journal of Image and Graphics.

• 2019, Second prize of Facebook Hack-a-Project Spring (2 out of 12 teams).

Master Project: Exploring the relation between cancers and mRNA with computer science methods

• Aim to find out which combination of mRNA bases have information of cancer metastasis.
• Built a complex network based on correlations among mRNAs using Python and Gephi.
• Did network analysis to filter out informative mRNAs.
• Currently researching on more effective network analysis and machine learning techniques,
• And how to predict cancers based on the filtered mRNAs with ML models.

Paid summer internship at UCL computer science AI research group

• A simulation study on Twitter breaking news spreading based on ML Models.
• Analysed, reproduced and evaluated previous ML models and methods.
• Will do data mining, feature selection and adding textural features(NLP) to the existing model.

Forecasting influenza-like illness rates using online search data

• Use Google query logs data to train a model to nowcast and forecast influenza-like illness.
• Examined state-of-the-art approaches, implemented and compared different ML models.
• Built a super-ensemble model which has a high performance.

Summer Research on Deep Learning & Neural Network Data Augment

• Use pytorch to implement CNN and research on current data augmentation approaches.
• Merge 4 data augment strategies (Cutout, Mixup, Cutmix and Autoaugment) into one project code.
• Train the RNN with different data augmentation strategies on different Dataset x RNN models.
• Analyse and contrast on the experiment results.
• Come up with a new strategy called Random-mix and run experiments to test the capability.

How to Change the World–Cities

• Designed a season responsive solution: Transportable Respiratory Environmental Equipment. (T.R.E.E.)
• Aims to reduce air pollution during Medellin’s smog seasons alongside the government air quality improvement plan.
• It is a modular system consisting of air filtering, mist spreading and an interactive information display.

Microsoft AI/Deep learning with Cortana Project

• Using Ionic, Microsoft Speech API, Google map API, Cognitive Service, Azure and Bot framework.
• Invoke Cortana and use voice commands to interact with Soundscape App.
• Incorporate AI to enable Cortana to learn from people’s behaviour and make recommendations. Integrate a chatbot to enable people to interact with Soundscape and learn to use it effectively.

Great Ormond Street Hospital 3D VR tour Project

• Using Unity and google VR SDK to develop an Android 3D VR tour Application to demo a new modern hospital concept. People can wear google VR glasses to see a further modern hospital.
• Users can interact with the app by looking at a specific item in the VR.

UCL Computer Science News Portal Project

• Team leader, front-end developer, back-end developer and researcher. Using Azure, PHP, HTML, CSS, JavaScript, MySQL, jQuery, Bootstrap and Github.
• Created an online portal for students to upload media on. Admin can login to manage, delete, download or tweet an article.
• pH level, temperature and stirring speed, all of which are integrated into one user interface.

Global Health Care-bioreactor controlling Project

• Team leader of the heating sub-system.
• A bioreactor is built to produce the vaccines. Worked on the controls of pH level, temperature and stirring speed, all of which integrated into one user