Picture for Xiang Xiang

Xiang Xiang

Lightweight Human Pose Estimation Using Heatmap-Weighting Loss

May 21, 2022
Shiqi Li, Xiang Xiang

Figure 1 for Lightweight Human Pose Estimation Using Heatmap-Weighting Loss

Figure 2 for Lightweight Human Pose Estimation Using Heatmap-Weighting Loss

Figure 3 for Lightweight Human Pose Estimation Using Heatmap-Weighting Loss

Figure 4 for Lightweight Human Pose Estimation Using Heatmap-Weighting Loss

* 7 pages, 5 figures

Access Paper or Ask Questions

Mapping Emulation for Knowledge Distillation

May 21, 2022
Jing Ma, Xiang Xiang, Zihan Zhang, Yuwen Tan, Yiming Wan, Zhigang Zeng, Dacheng Tao

Figure 1 for Mapping Emulation for Knowledge Distillation

Figure 2 for Mapping Emulation for Knowledge Distillation

Figure 3 for Mapping Emulation for Knowledge Distillation

Figure 4 for Mapping Emulation for Knowledge Distillation

* 11 pages, 5 figures, 3 lemmas, 3 corollaries, 1 algorithm

Access Paper or Ask Questions

Long-Tailed Classification with Gradual Balanced Loss and Adaptive Feature Generation

Feb 28, 2022
Zihan Zhang, Xiang Xiang

Figure 1 for Long-Tailed Classification with Gradual Balanced Loss and Adaptive Feature Generation

Figure 2 for Long-Tailed Classification with Gradual Balanced Loss and Adaptive Feature Generation

Figure 3 for Long-Tailed Classification with Gradual Balanced Loss and Adaptive Feature Generation

Figure 4 for Long-Tailed Classification with Gradual Balanced Loss and Adaptive Feature Generation

Access Paper or Ask Questions

Adversarial Deep Structured Nets for Mass Segmentation from Mammograms

Dec 25, 2017
Wentao Zhu, Xiang Xiang, Trac D. Tran, Gregory D. Hager, Xiaohui Xie

Figure 1 for Adversarial Deep Structured Nets for Mass Segmentation from Mammograms

Figure 2 for Adversarial Deep Structured Nets for Mass Segmentation from Mammograms

Figure 3 for Adversarial Deep Structured Nets for Mass Segmentation from Mammograms

Figure 4 for Adversarial Deep Structured Nets for Mass Segmentation from Mammograms

* Accepted by ISBI2018. arXiv admin note: substantial text overlap with arXiv:1612.05970

Access Paper or Ask Questions