Research Interests

My research interests lie in the areas of computer vision, applied deep learning, and machine learning.

Research Projects

Refining and Generalizing Trained CNNs

• Presenting refinement methods to mitigate overfitting and reduce the number of parameters.

Utilizing Mutation Test to Evaluate the Robustness of CNNs

• Presenting D-Score to assess CNNs’ robustness and fitness to specific datasets through mutation operators.
• Proposing a D-Score guided data augment method to improve CNNs’ robustness with respect to feature shifting.

Vision-based Depth Estimation from Monocular Single Image

• Presented an attention mechanism-based deep learning network and a novel loss function.
• Effectively generated depth maps for each pixel from only one image.
• Extensive experiments showed that it beats most SOTA methods, published a paper at a conference.

Multi-label Classification on Images with Missing Labels

• Proposed several approaches to address the problem of classification with partially-labeled data.
• Alleviating label imbalance, especially in large labeling spaces.
• Implemented by Python and PyTorch and experimented on several large-scale image datasets.

Quantum Programming and Computing

• Investigated IBM QISKit extensively, and existing quantum programming methods and algorithms.
• Proposed an effective compilation method of quantum programs, which only uses 74.7\% of the logic gates of the IBM standard method and the efficiency is increased by 7.75 times.
• Published one research paper at a conference and translated one book from English to Chinese.