3D Shape Classification——Point-based Discrete Convolution Network
點雲分類——離散卷積 Pointwise CNN PointCNN 卷積核權重不是分配給點雲，而是分配給網格，落在同一網格中的點獲得相同的權重。 代表1： Pointwise CNN 卷積核在整個點雲空間滑動，逐點進行卷積。對於落在同一格子的點，取他們的平均值與w相乘。關鍵在於如何分配w。 上圖把中心點附近的區域分成柵格，每個柵格內的特徵先取平均，然後再乘以柵格內的卷積權重，得到一個柵格的
相關文章
- 1. 3D Shape Classification——Point-based Continuous Convolution Networks
- 2. 3D Shape Classification——Point-based Graph based
- 3. 論文筆記《Active Convolution: Learning the Shape of Convolution for Image Classification》
- 4. 【論文筆記】Active Convolution: Learning the Shape of Convolution for Image Classification
- 5. 3D Shape Classification——Point-based Data Indexing based
- 6. Graphic Convolution Network(GCN)
- 7. Paper reading：PointGrid： A Deep Network for 3D Shape Understanding
- 8. 3D Machine Learning Single Object Classification
- 9. convolution backbone network——mobilenetv3
- 10. Transposed convolution & Fully Convolutional Neural Network
- 更多相關文章...
- • Lua 面向對象 - Lua 教程
- • C# 多態性 - C#教程
- • 互聯網組織的未來：剖析GitHub員工的任性之源
- • Flink 數據傳輸及反壓詳解
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. 3D Shape Classification——Point-based Continuous Convolution Networks
- 2. 3D Shape Classification——Point-based Graph based
- 3. 論文筆記《Active Convolution: Learning the Shape of Convolution for Image Classification》
- 4. 【論文筆記】Active Convolution: Learning the Shape of Convolution for Image Classification
- 5. 3D Shape Classification——Point-based Data Indexing based
- 6. Graphic Convolution Network(GCN)
- 7. Paper reading：PointGrid： A Deep Network for 3D Shape Understanding
- 8. 3D Machine Learning Single Object Classification
- 9. convolution backbone network——mobilenetv3
- 10. Transposed convolution & Fully Convolutional Neural Network