{"provider_url":"https://hatena.blog","provider_name":"Hatena Blog","blog_title":"Ridge-institute R&D Blog","author_url":"https://blog.hatena.ne.jp/Ridge-i_obarads/","type":"rich","published":"2022-08-19 15:47:25","description":"\u3053\u3093\u306b\u3061\u306f\uff0cRidge-i\u306e@obarads\u3067\u3059\uff0e\u672c\u8a18\u4e8b\u3067\u306f\uff0cImage2Point: 3D Point-Cloud Understanding with Pretrained 2D ConvNets [Xu et al, 2021]\u3092\u7d39\u4ecb\u3057\u307e\u3059(\u3053\u3061\u3089\u306e\u8ad6\u6587\u3067\u306fNeural Collapse\u306b\u3064\u3044\u3066\u3082\u53d6\u308a\u6271\u3063\u3066\u3044\u307e\u3059\u304c\uff0c\u9577\u304f\u306a\u308b\u305f\u3081\u4eca\u56de\u306f\u7701\u304d\u307e\u3059)\uff0e arxiv.org \u306f\u3058\u3081\u306b \u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u3084\u81ea\u52d5\u904b\u8ee2\u306a\u3069\u306e\u5206\u91ce\u3067\u306f\uff0c\u70b9\u7fa4\u3068\u547c\u3070\u308c\u308b\u30c7\u30fc\u30bf\u304c\u6271\u308f\u308c\u308b\u3053\u3068\u304c\u3042\u308a\u307e\u3059\uff0e\u70b9\u7fa4\u306f\uff0cLiDAR\u3092\u306f\u3058\u3081\u3068\u3059\u308b\u6df1\u5ea6\u30bb\u30f3\u30b5\u30fc\u304b\u3089\u53d6\u5f97\u3059\u308b\u3053\u3068\u304c\u53ef\u80fd\u3067\u3042\u308a\uff0c\u53d6\u5f97\u3055\u308c\u305f\u70b9\u7fa4\u306f3D\u30de\u30c3\u30d7\u3084\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\u306a\u3069\u306b\u5229\u7528\u2026","image_url":"https://cdn-ak.f.st-hatena.com/images/fotolife/R/Ridge-i_obarads/20220815/20220815161108.png","categories":["point cloud"],"blog_url":"https://iblog.ridge-i.com/","version":"1.0","height":"190","url":"https://iblog.ridge-i.com/entry/2022/08/19/154725","author_name":"Ridge-i_obarads","title":"Image2Point: 3D Point-Cloud Understanding with Pretrained 2D ConvNets","width":"100%","html":""}