A5101506232- Advanced Vision and Imaging
- Face recognition and analysis
- Crop Yield and Soil Fertility
- Wireless Sensor Networks for Data Analysis
- Remote-Sensing Image Classification
- Remote Sensing and LiDAR Applications
- Generative Adversarial Networks and Image Synthesis
- Machine Learning and ELM
- Advanced Image Processing Techniques
- Optical measurement and interference techniques
- Domain Adaptation and Few-Shot Learning
Northeastern University
2018-2024
Northeastern University
2019
Tianjin Academy of Agricultural Sciences
2012
Enhancing target domain discriminability is a key focus in Unsupervised Domain Adaptation (UDA) for HyperSpectral Image (HSI) classification. However, existing methods overlook bringing similar cross-domain samples closer together the feature space to achieve indirect transfer of source classification knowledge. To overcome this issue, we propose Multi-Task Learning-based (MTLDA) method. MTLDA incorporates an inductive mechanism into adversarial training, transferring knowledge...
We propose an image‐based face swapping algorithm, which can be used to replace the in reference image with same facial shape and features as input face. First, a alignment is made based on group of detected landmarks, so that aligned are consistent size posture. Secondly, warping algorithm triangulation presented adjust its background according faces. In order achieve more accurate swapping, parsing introduced realize detection face‐ROIs, then face‐ROI replaced face‐ROI. Finally, Poisson...
Convolutional autoencoders are making a significant impact on computer vision and signal processing communities. In this work, convolutional autoencoder denoising method is proposed to restore the corrupted laser stripe images of depth sensor, which directly reduces external noise sensor so as increase its accuracy. To reduce amount training data avoid overfitting, patch size image determined, basis small-scale dataset called Laser Stripe Image Patch (LSIP) created. Also, 14-layers...