{"id":1393,"date":"2017-09-01T12:03:19","date_gmt":"2017-09-01T03:03:19","guid":{"rendered":"https:\/\/d-itlab.co.jp\/?p=1393"},"modified":"2026-03-10T12:03:39","modified_gmt":"2026-03-10T03:03:39","slug":"20170901","status":"publish","type":"post","link":"https:\/\/d-itlab.co.jp\/en\/blog\/20170901\/","title":{"rendered":"BMVC2016\u51fa\u5f35\u5831\u544a"},"content":{"rendered":"

9\u670819\u65e5\u301c22\u65e5\u306b\u30a4\u30ae\u30ea\u30b9\u306eYork\u3067\u958b\u50ac\u3055\u308c\u305f\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30d3\u30b8\u30e7\u30f3\u306e\u5b66\u4f1aBMVC2016\u306b\uff0c\u6211\u3005\u306e\u7814\u7a76\uff08\u201dFast Eigen Matching\u201d\uff09\u306e\u767a\u8868\u3068\u8abf\u67fb\u306e\u305f\u3081\u53c2\u52a0\u3057\u3066\u304d\u307e\u3057\u305f\uff0e<\/p>\n

BMVC\u3078\u306e\u53c2\u52a0\u306f\u521d\u3081\u3066\u3067\u3057\u305f\u304c\uff0c\u30b3\u30f3\u30d1\u30af\u30c8\u306b\u307e\u3068\u307e\u3063\u305f\u4f1a\u8b70\u3067\u3057\u305f\uff0e\u500b\u4eba\u7684\u306b\u7279\u306b\u6c17\u306b\u306a\u3063\u305f\u767a\u8868\u3092\u5e7e\u3064\u304b\u7d39\u4ecb\u3057\u305f\u3044\u3068\u601d\u3044\u307e\u3059\uff0e
\n\u203b\u79c1\u306e\u524d\u63d0\u77e5\u8b58\u4e0d\u8db3\u3067\uff0c\u7406\u89e3\u304c\u66d6\u6627\u306a\u70b9\u3084\uff0c\u9593\u9055\u3063\u305f\u7406\u89e3\u3092\u7406\u89e3\u3092\u3057\u3066\u3044\u308b\u70b9\u304c\u3042\u308b\u3068\u601d\u3044\u307e\u3059\uff0e\u9593\u9055\u3044\u7b49\u3054\u3056\u3044\u307e\u3057\u305f\u3089\u3054\u6307\u6458\u3044\u305f\u3060\u3051\u307e\u3059\u3068\u5e78\u3044\u3067\u3059\uff0e<\/p>\n

BMVC<\/strong>
\nBMVC\u306f\uff0c1985\u5e74\u306b\u59cb\u307e\u3063\u305f\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30d3\u30b8\u30e7\u30f3(CV)\u306e\u5b66\u4f1a\u3067\u3059\uff0e
\n\u5b66\u4f1a\u540d\u306b\u306fBritish\u304c\u3064\u3044\u3066\u3044\u307e\u3059\u304c\uff0c\u591a\u304f\u306e\u8ad6\u6587\u304c\u30a4\u30ae\u30ea\u30b9\u56fd\u5916\u304b\u3089\u306e\u3082\u306e\u3067\uff0c\u65e5\u672c\u304b\u3089\u306e\u767a\u8868\u3082\u6570\u4ef6\u3042\u308a\u307e\u3057\u305f\uff0e
\n\u30aa\u30e9\u30fc\u30eb\u767a\u8868\u304c\u30b7\u30f3\u30b0\u30eb\u30c8\u30e9\u30c3\u30af\u306a\u306e\u3067\uff08\u7406\u89e3\u3067\u304d\u308b\u304b\u306f\u5225\u3068\u3057\u3066\uff09\u4e00\u4eba\u3067\u884c\u3063\u3066\u3082\u5168\u90e8\u8074\u8b1b\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\uff0e\u30dd\u30b9\u30bf\u30fc\u3082\u5168\u6587\u3067100\u4ef6\u7a0b\u5ea6\u3067\u3059\uff0e
\nMicrosoft Academic\u30e9\u30f3\u30ad\u30f3\u30b0\u3067\u306fCV\u5206\u91ce\u3067\u306f\uff14\u756a\u76ee\u3067\uff0c\u767a\u8868\u6570\u304c\u5c11\u306a\u3044\u5272\u306b\u306f\u5f71\u97ff\u529b\u306e\u3042\u308b\u8ad6\u6587\u304c\u591a\u3044\u3088\u3046\u306b\u611f\u3058\u307e\u3059\uff0e<\/p>\n

\u4eca\u5e74\u306e\u8ad6\u6587\u306fSupplemental Material\u3092\u542b\u3081\u3066\u516c\u5f0fHP<\/a>\u3088\u308a\u5168\u90e8\u95b2\u89a7\u3067\u304d\u307e\u3059\uff0e
\n\u6c17\u306b\u306a\u3063\u305f\u767a\u8868\u306e\u4e2d\u304b\u3089\u5e7e\u3064\u304b\u30d4\u30c3\u30af\u30a2\u30c3\u30d7\u3057\u3066\u7d39\u4ecb\u3057\u307e\u3059\uff0e<\/p>\n

\u79c1\u306e\u8aac\u660e\u3067\u306f\u89e6\u308a\u7a0b\u5ea6\u3057\u304b\u4f1d\u308f\u3089\u306a\u3044\u3068\u304a\u3082\u3044\u307e\u3059\uff0e\u5404\u8ad6\u6587\u6bce\u306b1\u30da\u30fc\u30b8\u306eExtended Abstract\u304c\u516c\u958b\u3055\u308c\u3066\u3044\u307e\u3059\u306e\u3067\uff0c\u305d\u308c\u3092\u773a\u3081\u308b\u306e\u3082\u3044\u3044\u3068\u601d\u3044\u307e\u3059\uff0e<\/p>\n

\n

[Tutorial] Towards Affordable Self-driving Cars<\/strong>
\n\u30c8\u30ed\u30f3\u30c8\u5927\u306eRaquel Urtasun\u3055\u3093\u306b\u3088\u308b\uff0c\u81ea\u52d5\u904b\u8ee2\u95a2\u9023\u306e\u6280\u8853\u306e\u95a2\u3059\u308b\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\uff0c\u7279\u306b\u5f7c\u5973\u305f\u3061\u306e\u7814\u7a76\u30b0\u30eb\u30fc\u30d7\u306e\u7814\u7a76\u6210\u679c\u306b\u3064\u3044\u3066\u8aac\u660e\u3057\u3066\u3044\u307e\u3057\u305f\uff0e
\nRaquel\u3055\u3093\u306e\u30b0\u30eb\u30fc\u30d7\u306f\uff0cKITTI\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u306a\u3069\u81ea\u52d5\u904b\u8ee2\u306b\u95a2\u3059\u308b\u7814\u7a76\u3092\uff11\u3064\u306e\u67f1\u3068\u3057\u3066\u3044\u3066\uff0cCVPR,NIPS\u306a\u3069\u306e\u30c8\u30c3\u30d7\u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u306b\u6bce\u5e74\u305f\u304f\u3055\u3093\u306e\u8ad6\u6587\u3092\u901a\u3057\u3066\u3044\u307e\u3059\uff0e<\/p>\n

\u4eca\u56de\u306e\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\u306f\uff0c\u30bf\u30a4\u30c8\u30eb\u306bAffordable\u3068\u3042\u308b\u3088\u3046\u306b\uff0c\u30d9\u30ed\u30c0\u30a4\u30f3\u306a\u3069\u3092\u7528\u3044\u305a\u306b\u30ab\u30e1\u30e9\u306a\u3069\u65e2\u5b58\u306e\u8eca\u8f09\u30bb\u30f3\u30b7\u30f3\u30b0\u30c7\u30d0\u30a4\u30b9\u306b\u8fd1\u3044\u69cb\u6210\u3067\u81ea\u52d5\u904b\u8ee2\u3092\u5b9f\u73fe\u3059\u308b\u305f\u3081\u306e\u7814\u7a76\u306b\u3064\u3044\u3066\u306e\u767a\u8868\u3057\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

\u8eca\u8f09\u5358\u773c\u30ab\u30e1\u30e9\u753b\u50cf\u306e\u30bb\u30de\u30f3\u30c6\u30a3\u30c3\u30af\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\uff0cOpenStreetMap\uff08OSM\uff09\u306e\u30ea\u30f3\u30af\u30c7\u30fc\u30bf\u3068Visual Oddmetory\u3092\u7528\u3044\u305f\u81ea\u5df1\u4f4d\u7f6e\u63a8\u5b9a\uff0cOSM\u3092\u7528\u3044\u305f\u8eca\u8f09\u30ab\u30e1\u30e9\u753b\u50cf\u3078\u306e\u30e9\u30d9\u30eb\u4ed8\u6280\u8853\u306a\u3069\u304c\u7d39\u4ecb\u3055\u308c\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

OSM\u3068Visual Oddmetory\u3092\u7528\u3044\u305f\u81ea\u5df1\u4f4d\u7f6e\u63a8\u5b9a\u3067\u306f\uff0c\u6570\u30ad\u30ed\uff58\u6570\u30ad\u30ed\u56db\u65b9\u306e\u5e83\u7bc4\u56f2\u30a8\u30ea\u30a2\u304b\u3089Visual Oddmetory\u306e\u60c5\u5831\u3092OSM\u306e\u30ea\u30f3\u30af\u60c5\u5831\u3068\u30b0\u30e9\u30d5\u3092\u7528\u3044\u3066\u30de\u30c3\u30c1\u30f3\u30b0\u3059\u308b\u3053\u3068\u3067\uff0c\u52b9\u7387\u7684\u306b\u73fe\u5728\u4f4d\u7f6e\u306e\u63a8\u5b9a\u3092\u884c\u3046\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3057\u305f\uff0e\u554f\u984c\u8a2d\u5b9a\u3084\u539f\u7406\u304c\u3068\u3066\u3082\u9762\u767d\u3044\u3068\u601d\u3044\u307e\u3057\u305f\uff0e\u3053\u3053<\/a>\u306b\u30d3\u30c7\u30aa\u304c\u516c\u958b\u3055\u308c\u3066\u3044\u307e\u3059\uff0e<\/p>\n

Adding Synchronization and Rolling Shutter in Multi-Camera Bundle Adjustment<\/strong>
\n\u591a\u304f\u306e\u30b3\u30f3\u30b9\u30fc\u30de\u5411\u3051\u30ab\u30e1\u30e9\u306f\uff0cRolling Shutter\u3068\u547c\u3070\u308c\u308b\u753b\u7d20\u3054\u3068\u306b\u8f1d\u5ea6\u5024\u304c\u30b7\u30fc\u30b1\u30f3\u30b7\u30e3\u30eb\u306b\u8aad\u307f\u8fbc\u307e\u308c\u308b\u69cb\u6210\u306b\u306a\u3063\u3066\u3044\u307e\u3059\uff0e
\n\u3064\u307e\u308a\uff0c\u753b\u7d20\u306e\u5de6\u4e0a\u3068\u53f3\u4e0b\u3067\u306f\uff0c\u6642\u9593\u7684\u306b\u7570\u306a\u308b\u30bf\u30a4\u30df\u30f3\u30b0\u306e\u8f1d\u5ea6\u60c5\u5831\u304c\u5f97\u3089\u308c\u308b\u3053\u3068\u306b\u306a\u308a\u307e\u3059\uff0e
\n\u5f7c\u3089\u306f\uff0c\u30de\u30eb\u30c1\u30ab\u30e1\u30e9\u306eBundle Adjustment\u306b\u304a\u3044\u3066\uff0c\u30ab\u30e1\u30e9\u9593\u306eRolling Shutter\u306e\u540c\u671f\u305a\u308c\u3082\u30d1\u30e9\u30e1\u30fc\u30bf\u3068\u3057\u3066\u52a0\u3048\u305f\u5b9a\u5f0f\u5316\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e
\nGoPro\u30ab\u30e1\u30e9\u3092\uff14\u3064\u7528\u3044\u305f\u5168\u5468\u30ab\u30e1\u30e9\u3092\u5236\u4f5c\u3057\uff0c\u305d\u308c\u3092\u7528\u3044\u305fSLAM\u306e\u7d50\u679c\u3092\u7d39\u4ecb\u3057\u3066\u3044\u307e\u3057\u305f\uff0e\u767a\u8868\u3067\u306fLoop Clousure\u51e6\u7406\u306a\u3057\u3067\u3082\u30c9\u30ea\u30d5\u30c8\u304c\u5c11\u306a\u3044\u7d50\u679c\u304c\u5f97\u3089\u308c\u308b\u3053\u3068\u3092\u30a2\u30d4\u30fc\u30eb\u3057\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

Event Camera\u3092\u7528\u3044\u305f\u7814\u7a76<\/strong>
\n\u4eca\u56de\uff0cEvent Camera\u3092\u7528\u3044\u305f\u767a\u8868\u304c\uff13\u4ef6\u3042\u308a\u307e\u3057\u305f\uff0e
\nEMVS: Event-based Multi-View Stereo, Real-Time Intensity-Image Reconstruction for Event Cameras Using Manifold Regularisation, Event-Based Hough Transform in a Spiking Neural Network for Multiple Line Detection and Tracking Using a Dynamic Vision Sensor
\nEvent Camera\u306f\uff0c\u901a\u5e38\u306e\u5168\u753b\u7d20\u3092\u30d5\u30ec\u30fc\u30e0\u6bce\u306b\u51fa\u529b\u30ab\u30e1\u30e9\u3068\u306f\u7570\u306a\u308a\uff0cEvent(\u8f1d\u5ea6\u5024\u306e\u5909\u5316)\u304c\u3042\u3063\u305f\u753b\u7d20\u306e\u60c5\u5831\u306e\u307f\u51fa\u529b\u3059\u308b\u30ab\u30e1\u30e9\u3067\u3059\uff0e
\n\u5177\u4f53\u7684\u306f\uff0c\uff08x\u5ea7\u6a19\uff0cy\u5ea7\u6a19\uff0c\u30bf\u30a4\u30e0\u30b9\u30bf\u30f3\u30d7\uff0c\u6975\u6027\uff09\u3092\u30a4\u30d9\u30f3\u30c8\u60c5\u5831\u3068\u3057\u3066\u4f1d\u3048\u307e\u3059\uff0e\u6975\u6027\u306f\uff0c\u8f1d\u5ea6\u304c\u95be\u5024\u3088\u308a\u6e1b\u3063\u305f\u304b\u5897\u3048\u305f\u3092\u793a\u3059\u30d0\u30a4\u30ca\u30ea\u5024\u3067\u3059\uff0e
\n\u6642\u9593\u89e3\u50cf\u5ea6\u306f10\u30de\u30a4\u30af\u30ed\u79d2\u3068\u975e\u5e38\u306b\u77ed\u304f\uff0c\u8f1d\u5ea6\u5909\u5316\u306f\u7dda\u5f62\u3067\u306a\u304fLog\u30b9\u30b1\u30fc\u30eb\u3067\u691c\u51fa\u3055\u308c\u308b\u305f\u3081\u30c0\u30a4\u30ca\u30df\u30c3\u30af\u30ec\u30f3\u30b8\u304c\u975e\u5e38\u306b\u5e83\u304f\u6697\u3044\u5834\u6240\u3068\u660e\u308b\u3044\u5834\u6240\u304c\u6df7\u5728\u3059\u308b\u74b0\u5883\u3067\u3082\u3061\u3083\u3093\u3068\u8f1d\u5ea6\u5909\u5316\u3092\u3068\u3089\u3048\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\uff0e
\n\u3044\u308f\u3086\u308bbio-inspired vision sensors\u3068\u3044\u3046\u3082\u306e\u3067\uff0c\u751f\u7269\u306eVision\u30b7\u30b9\u30c6\u30e0\u306b\u8fd1\u3044\u69cb\u6210\u3068\u306a\u3063\u3066\u3044\u307e\u3059\uff0e<\/p>\n

\u4e0a\u8a18\u306e3\u3064\u306e\u767a\u8868\u306f\u3044\u305a\u308c\u3082inilabs\u3068\u3044\u3046\u4f1a\u793e\u306eDVS128<\/a>\u3068\u3044\u3046\u30ab\u30e1\u30e9\u3092\u7528\u3044\u3066\u3044\u307e\u3044\u305f\uff0e<\/p>\n

EMVS: Event-based Multi-View Stereo<\/strong>\u3067\u306f\uff0c\u30ab\u30e1\u30e9\u30e2\u30fc\u30b7\u30e7\u30f3\u304c\u65e2\u77e5\u306e\u524d\u63d0\u3067\uff0cEvent Camera\u304b\u3089\u5f97\u3089\u308c\u308b\u30a8\u30c3\u30b8\u60c5\u5831\u30923\u6b21\u5143\u30dc\u30af\u30bb\u30eb\u4e0a\u306bVoting\u3059\u308b\u3053\u3068\u3067\u5bfe\u5fdc\u70b9\u63a2\u7d22\u306a\u3057\u306b\uff0c3\u6b21\u5143\u5fa9\u5143\u3092\u884c\u3046\u3082\u306e\u3067\u3059\uff0e\u6570\u884c\u3067\u5b9f\u88c5\u3067\u304d\u308b\u3068\u3066\u3082\u30b7\u30f3\u30d7\u30eb\u306a\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3067\u3059\u304c\uff0c\u5fa9\u5143\u7d50\u679c\u306f\u3051\u3063\u3053\u3046\u30ad\u30ec\u30a4\u3067\u9762\u767d\u3044\u30a2\u30a4\u30c7\u30a3\u30a2\u3060\u3068\u611f\u3058\u307e\u3057\u305f\uff0e<\/p>\n

Real-Time Intensity-Image Reconstruction for Event Cameras Using Manifold Regularisation\u306f\uff0c\u30b9\u30d1\u30fc\u30b9\u304b\u3064\u30d0\u30a4\u30ca\u30ea\u306eEvent\u60c5\u5831\u304b\u3089\uff0c\u30c7\u30f3\u30b9\u3067\u968e\u8abf\u3092\u6301\u3063\u305f\u8f1d\u5ea6\u753b\u50cf\u3092\u5fa9\u5143\u3059\u308b\u6280\u8853\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e<\/p>\n

Event-Based Hough Transform in a Spiking Neural Network for Multiple Line Detection and Tracking Using a Dynamic Vision Sensor\u306f\uff0cHought\u7a7a\u9593\u306b\u5c04\u5f71\u3057\u305fEvent\u60c5\u5831\u304b\u3089Spiking Neural Network\u3092\u7528\u3044\u3066\u76f4\u7dda\u691c\u51fa\u3068\u30c8\u30e9\u30c3\u30ad\u30f3\u30b0\u3092\u884c\u3063\u3066\u3044\u3066\uff0c\u5168\u51e6\u7406\u3092FPGA\u4e0a\u3067\u5b9f\u88c5\u3057\u3066\u3044\u308b\u3088\u3046\u3067\u3059\uff0e<\/p>\n

Global Deconvolutional Networks for Semantic Segmentation <\/strong>
\nFully Convolutional Network(FCN)\u3092\u7528\u3044\u305fSemantic Segmentation\u306e\u6539\u826f\u30a2\u30eb\u30b4\u3067\u3059\uff0eFCN\u3067\u306f\u30a8\u30f3\u30b3\u30fc\u30c9\u3055\u308c\u305f\u5c0f\u3055\u306a\u753b\u50cf$x$\u304b\u3089\uff0c\u8fd1\u508d\u306e\u30d4\u30af\u30bb\u30eb\u306e\u88dc\u9593\uff08\u62e1\u5927\uff09\u306b\u3088\u308a\u6700\u7d42\u7684\u306aSemanticSegmentation\u306e\u7d50\u679c\u3092\u51fa\u529b\u3057\u3066\u3044\u307e\u3059\uff0e\u672c\u7814\u7a76\u3067\u306f\u3053\u306e$x$\u306e\u62e1\u5927\u51e6\u7406\u3092\uff0c$K_hxK^{top}_w$\u306e\u3088\u3046\u306b\u30b0\u30ed\u30fc\u30d0\u30eb\u306a\u76f8\u95a2\u3092\u8003\u616e\u3057\u305f\u62e1\u5927\u30d1\u30e9\u30e1\u30fc\u30bf\u3092\u5b66\u7fd2\u3059\u308b\u3053\u3068\u3067\u6027\u80fd\u3092\u5411\u4e0a\u3057\u3066\u3044\u307e\u3059\uff0e<\/p>\n

Graph Based Convolutional Neural Network <\/strong>
\n\u30b0\u30e9\u30d5\u4fe1\u53f7\u51e6\u7406\u306e\u539f\u7406\u3092\u7528\u3044\u305f\u30b0\u30e9\u30d5\u4e0a\u3067\u5b9a\u7fa9\u3055\u308c\u308bNeural Net(NN)\u3067\u3059\uff0e\u30b0\u30e9\u30d5\u4e0a\u306e\u7573\u307f\u8fbc\u307f\u3084\u30d7\u30fc\u30ea\u30f3\u30b0\uff0c\u3053\u308c\u3089\u306e\u6f14\u7b97\u306b\u5bfe\u3059\u308b\u8aa4\u5dee\u306e\u9006\u4f1d\u642c\u624b\u6cd5\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e
\nmnist\u30c7\u30fc\u30bf\u3092irregular\u30b0\u30ea\u30c3\u30c9\u306b\u5909\u63db\u3057\u305f\u30c7\u30fc\u30bf\u3092\u7528\u3044\u3066\u63d0\u6848\u6cd5\u306e\u6709\u7528\u6027\u3092\u8a55\u4fa1\u3057\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

Improving Weakly-Supervised Object Localization By Micro-Annotation<\/strong>
\n\u753b\u50cf\u5168\u4f53\u3078\u306e\u30e9\u30d9\u30eb\u60c5\u5831\u3060\u3051\u304c\u4ed8\u4e0e\u3055\u308c\u305f\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u7528\u3044\u3066\uff0c\u7269\u4f53\u306e\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\u3092\u884c\u3046Weakly-Supervised Object Localization\u306e\u6027\u80fd\u5411\u4e0a\u306b\u95a2\u3059\u308b\u7814\u7a76\u3067\u3059\uff0e
\n\u4f8b\u3048\u3070\uff0c\u30df\u30c4\u30d0\u30c1\u3068\u82b1\u306a\u3069\u540c\u6642\u306b\u767a\u751f\u3057\u6613\u3044\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u306e\u5834\u5408\uff0cWeakly-Supervised\u30c7\u30fc\u30bf\u304b\u3089\u306f\u305d\u308c\u3089\u3092\u5206\u96e2\u3059\u308b\u60c5\u5831\u304c\u306a\u3044\u306e\u3067\uff0c\u6027\u80fd\u304c\u60aa\u304f\u306a\u3063\u3066\u3057\u307e\u3044\u307e\u3059\uff0e
\n\u5f7c\u3089\u306f\uff0c\u7269\u4f53\u8a8d\u8b58\u306e\u5b66\u7fd2\u3092\u884c\u3063\u305f\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u4e2d\u9593\u5c64\u304b\u3089\u5f97\u3089\u308c\u308bLocalization Socre map\u304c\u753b\u9762\u306b\u8868\u793a\u3055\u308c\u3066\u3044\u308b\u30e9\u30d9\u30eb\u306b\u5bfe\u5fdc\u3059\u308b\u9818\u57df\u3092\u793a\u3057\u3066\u3044\u308b\u304b\/\u5426\u304b\u3068\u3044\u3046Micro-Annotation\u60c5\u5831\u3092\u8ffd\u52a0\u3059\u308b\u3053\u3068\u3067\uff0cObject Localization\u306e\u6027\u80fd\u3092\u5411\u4e0a\u3059\u308b\u624b\u6cd5\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e
\nMicro-Annotation\u306f\uff0c\u30d0\u30a4\u30ca\u30ea\u306e\u30e9\u30d9\u30eb\u4ed8\u306a\u306e\u3067\u4f5c\u696d\u8ca0\u8377\u304c\u4f4e\u304f\uff0c\u5c11\u306a\u3044\u8ffd\u52a0\u4f5c\u696d\u3067\u6027\u80fd\u3092\u5411\u4e0a\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3053\u3068\u3092\u30a2\u30d4\u30fc\u30eb\u3057\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

Learning Neural Network Architectures using Backpropagation <\/strong>
\nNN\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u69cb\u6210\uff08\u5404\u5c64\u306eWidht\u3068\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u6df1\u3055\uff09\u81ea\u4f53\u3092NN\u3067\u5b66\u7fd2\u3057\u3088\u3046\u3068\u3044\u3046\u7814\u7a76\u3067\u3059\uff0e
\n\u91cd\u307f\u30d1\u30e9\u30e1\u30fc\u30bf\u306e\u6709\u52b9\u30fb\u7121\u52b9\u3092\u6c7a\u3081\u308b\u30d0\u30a4\u30ca\u30ea\u91cd\u307f\u306e\u5c0e\u5165\u306b\u3088\u308aWidht\u3092\u8868\u73fe\u3057\uff0c\u30d1\u30e9\u30e1\u30fc\u30bf\u306b\u3088\u3063\u3066\u306f\u7dda\u5f62\u5909\u63db\u3092\u8868\u73fe\u53ef\u80fd\u306aNonlinearity(tsReLU)\u3092\u5c0e\u5165\u3059\u308b\u3053\u3068\u3067\u6df1\u3055\u3092\u8868\u73fe\u3057\u3066\u3044\u307e\u3059\uff0e
\n\u3053\u308c\u3089\u306e\u30d1\u30e9\u30e1\u30fc\u30bf\u3092\u8aa4\u5dee\u9006\u4f1d\u642c\u306b\u3088\u308a\u6700\u9069\u5316\u3059\u308b\u3053\u3068\u3067\uff0c\u9069\u5207\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u69cb\u6210\u3092\u8a08\u7b97\u3057\u307e\u3059\uff0e<\/p>\n

OnionNet: Sharing Features in Cascaded Deep Classifiers <\/strong>
\n\u8907\u6570\u306eCNN\u3092\u30ab\u30b9\u30b1\u30fc\u30c9\u63a5\u7d9a\u3059\u308b\u3053\u3068\u3067\u8b58\u5225\u6027\u80fd\u3092\u5411\u4e0a\u3057\u305f\u624b\u6cd5\u3067\u3059\uff0e
\n\u500b\u3005\u306eCNN\u306f\u72ec\u7acb\u306b\u5b66\u7fd2\u3059\u308b\u306e\u3067\u306f\u306a\u304f\uff0c\u30d1\u30e9\u30e1\u30fc\u30bf\u3092\u5171\u6709\u3059\u308b\u3088\u3046\u306b\u3059\u308b\u3053\u3068\u3067\u30d1\u30e9\u30e1\u30fc\u30bf\u6570\u3092\u524a\u6e1b\u3057\u3066\u3044\u307e\u3059\uff0e<\/p>\n

Wide Residual Networks <\/strong>
\n\u4eca\u5e74\u306eBMVC\u3067\u4e00\u756a\u8a71\u984c\u306b\u306a\u3063\u305f\u8ad6\u6587\u304b\u3082\u3057\u308c\u307e\u305b\u3093\uff0e
\n\u975e\u5e38\u306bDeep\u306aResNet\u3068\u6bd4\u3079\u3066\uff0c\u5b9f\u306fWide\u3067Shallow\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u65b9\u304c\u6027\u80fd\u304c\u826f\u3044\u3068\u3044\u3046\u7d50\u679c\u3092\u793a\u3057\u305f\u8ad6\u6587\u3067\u3059\uff0e
\n\u901a\u5e38ResNet\u3067\u306f\uff0c3\u00d73\u306e\u7573\u307f\u8fbc\u307f\u5c64\u306e\u524d\u306b1\u00d71\u306e\u6b21\u5143\u5727\u7e2e\u3092\u884c\u3046\u5c64\u3092\u5c0e\u5165\u3059\u308b\u3053\u3068\u3067\uff0c\u591a\u5c64\u306b\u3057\u305f\u3068\u304d\u306eChannel\u6570\u3092\u6291\u5236\u3057\uff0c\u975e\u5e38\u306bDeep\u306a\u69cb\u6210\u3092\u5b9f\u73fe\u3057\u3066\u3044\u307e\u3059\uff0e
\nWide Residual Networks\u3067\u306f\uff0c\u3053\u306e\u6b21\u5143\u5727\u7e2e\u3092\u884c\u308f\u305a\u306b\u4e2d\u9593\u5c64\u306e\u30c1\u30e3\u30cd\u30eb\u306f\u6bd4\u8f03\u7684Wide\u306b\u3057\u3066\u305d\u306e\u4ee3\u308f\u308a\u306b\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u6df1\u3055\u3092Modelate\u306a\u6df1\u3055\u306b\u3057\u3066\u3044\u307e\u3059\uff0e\u305d\u3057\u3066\uff0c\u3053\u306e\u69cb\u6210\u306e\u65b9\u304c\u6027\u80fd\u304c\u826f\u3044\u3068\u3044\u3046\u5b9f\u9a13\u7d50\u679c\u3092\u793a\u3057\u3066\u3044\u307e\u3059\uff0e
\n\u4eca\u5e74\u306eImage Net\u30b3\u30f3\u30da\u3067\u3082\u3053\u308c\u306b\u985e\u4f3c\u3057\u305f\u69cb\u6210\u304c\u9ad8\u3044\u6027\u80fd\u3092\u793a\u3057\u3066\u3044\u308b\u3088\u3046\u3067\u3059\uff0e<\/p>\n

A Deep Primal-Dual Network for Guided Depth Super-Resolution <\/strong>
\n\u4f4e\u89e3\u50cf\u5ea6\u30c7\u30d7\u30b9\u3068\u9ad8\u89e3\u50cf\u5ea6RGB\u753b\u50cf\u3092\u5165\u529b\u3068\u3057\u3066\u9ad8\u89e3\u50cf\u5ea6\u30c7\u30d7\u30b9\u3092\u51fa\u529b\u3059\u308bCNN\u306e\u5f8c\u6bb5\u306b\uff0cCNN\u306e\u51fa\u529b\u3067\u3042\u308b\u9ad8\u89e3\u50cf\u5ea6\u30c7\u30d7\u30b9\u3092Total Valiation(TV)\u6b63\u5247\u5316\u3092\u7528\u3044\u3066\u30c7\u30ce\u30a4\u30b8\u30f3\u30b0\u3059\u308b\u30cd\u30c3\u30c8\u30ef\u30fc\u30af(primal-dual network(PDN))\u304c\u633f\u5165\u3055\u308c\u305f2\u3064\u306e\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u304b\u3089\u306a\u308b\u8d85\u89e3\u50cf\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u3067\u3059\uff0e
\n\u5f8c\u6bb5\u306ePDN\u306b\u3088\u308b\u6700\u9069\u5316\u3092\u9069\u5f53\u306aItereration\u56de\u6570\u5206\u3060\u3051Unlolling\u3059\u308b\u3053\u3068\u3067\uff0c\u6700\u7d42\u7684\u306a\u30a8\u30e9\u30fc\u3092CNN\u306e\u5c64\u307e\u3067\u9006\u4f1d\u642c\u3059\u308b\u624b\u6cd5\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e
\n\u305d\u306e\u307e\u307e\u3067\u306f\uff0c\u8aa4\u5dee\u9006\u4f1d\u642c\u3092\u4f7f\u3063\u3066CNN\u3068\u540c\u6642\u306b\u6700\u9069\u5316\u3067\u304d\u306a\u3044\u5c64\uff08PDN\uff09\u3092\uff0cUnrolling\u306b\u3088\u308a\u5c55\u958b\u3057\uff0cEnd-to-End\u3067\u8aa4\u5dee\u9006\u4f1d\u642c\u306b\u3088\u308b\u6700\u9069\u5316\u304c\u3067\u304d\u308b\u3088\u3046\u306b\u3057\u3066\u3044\u308b\u70b9\u304c\u975e\u5e38\u306b\u8208\u5473\u6df1\u3044\u3067\u3059\uff0e
\nUnlolling\u3092\u7528\u3044\u305f\u57fa\u672c\u7684\u306a\u30a2\u30a4\u30c7\u30a3\u30a2\u306f2016 ECCV\u306eATGV-Net(\u540c\u8457\u8005\u306b\u3088\u308b\u672c\u7814\u7a76\u306e\u5143\u8ad6\u6587)\u3067\u89e3\u8aac\u3055\u308c\u3066\u3044\u307e\u3059\uff0e<\/p>\n

Measuring the effect of nuisance variables on classifiers <\/strong>
\n\u591a\u30af\u30e9\u30b9\u8b58\u5225\u3092\u884c\u3046NN\u306e\u5916\u4e71\u306b\u5bfe\u3059\u308b\u30ed\u30d0\u30b9\u30c8\u6027\u3092\u8a55\u4fa1\u3059\u308b\u624b\u6cd5\u3092\u63d0\u6848\u3057\u305f\u7814\u7a76\u3067\u3059\uff0e
\n\u5165\u529b\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066$theta$\u3092\u4e8b\u524d\u5206\u5e03\u3068\u3059\u308b\u5909\u5f62$tau$\u3092\u884c\u3063\u305f\u3068\u304d\u306b\uff0c\u6761\u4ef6\u4ed8\u304dConfidence $p(l(x)|theta)$\u3092\u30b5\u30f3\u30d7\u30eb\u30ea\u30f3\u30b0\u306b\u3088\u308a\u8a08\u7b97\u3057\uff0c\u3053\u308c\u3092\u7528\u3044\u3066\u5916\u4e71$tau$\u306b\u5bfe\u3059\u308b\u30ed\u30d0\u30b9\u30c8\u6027$hat{rho_{tau}}$\u3092\u8a55\u4fa1\u3057\u307e\u3059\uff0e
\n\u8ad6\u6587\u306b\u306f\u69d8\u3005\u306a\u65e2\u5b58\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u5916\u4e71\uff08\u90e8\u5206\u7684\u306aAffine\u5909\u5f62\uff09\u306b\u5bfe\u3059\u308b\u5f71\u97ff\u3092\u63d0\u6848\u6cd5\u3067\u8a55\u4fa1\u3057\u305f\u7d50\u679c\u304c\u793a\u3055\u308c\u3066\u3044\u307e\u3059\uff0e
\n\u3053\u306e\u7814\u7a76\u4ee5\u5916\u306b\u3082\uff0cDNN\u306e\u6027\u80fd\u8a55\u4fa1\u306b\u95a2\u3059\u308b\u7814\u7a76\u304c\u3044\u304f\u3064\u304b\u767a\u8868\u3055\u308c\u3066\u3044\u307e\u3057\u305f\uff0e<\/p>\n

\n

York\u306b\u3064\u3044\u3066<\/strong>
\nYork\u306f\u3082\u3068\u3082\u3068\u30d0\u30a4\u30ad\u30f3\u30b0\u306e\u8857\u3060\u3063\u305f\u3088\u3046\u3067\u305d\u306e\u540d\u6b8b\u3067\uff0c\u57ce\u58c1\u3084\u57ce\u306e\u5f8c\u304c\u3042\u308a\uff0c\u98f2\u98df\u5e97\u3084\u304a\u571f\u7523\u5c4b\u3055\u3093\u304c\u3042\u308b\u8857\u306e\u4e2d\u5fc3\u90e8\u3082\u53e4\u3044\u30ec\u30f3\u30ac\u9020\u308a\u3067\u4e2d\u4e16\u306e\u96f0\u56f2\u6c17\u3067\u3059\uff0e
\n\u307e\u305f\u99c5\u306e\u8fd1\u304f\u306b\u56fd\u7acb\u9244\u9053\u535a\u7269\u9928\u304c\u3042\u308a\uff0c\u65b0\u5e79\u7dda\uff08\u3072\u304b\u308a\uff1f\uff09\u304c\u3042\u308a\u307e\u3057\u305f\uff0e<\/p>\n

\n
\n
\n\t\t\t\t\"\u65b0\u5e79\u7dda\"<\/a>\n\t\t\t<\/div>
\n\t\t\t\t\u9244\u9053\u535a\u7269\u9928
\n\t\t\t\t<\/figcaption><\/figure>\n
\n
\n\t\t\t\t\"\"<\/a>\n\t\t\t<\/div>\n<\/figure>\n
\n
\n\t\t\t\t\"\"<\/a>\n\t\t\t<\/div>\n<\/figure>\n
\n
\n\t\t\t\t\"\"<\/a>\n\t\t\t<\/div>\n<\/figure>\n
\n
\n\t\t\t\t\"\"<\/a>\n\t\t\t<\/div>\n<\/figure><\/div>\n","protected":false},"excerpt":{"rendered":"

9\u670819\u65e5\u301c22\u65e5\u306b\u30a4\u30ae\u30ea\u30b9\u306eYork\u3067\u958b\u50ac\u3055\u308c\u305f\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30d3\u30b8\u30e7\u30f3\u306e\u5b66\u4f1aBMVC2016\u306b\uff0c\u6211\u3005\u306e\u7814\u7a76\uff08\u201dFast Eigen Matching\u201d\uff09\u306e\u767a\u8868\u3068\u8abf\u67fb\u306e\u305f\u3081\u53c2\u52a0\u3057\u3066\u304d\u307e\u3057\u305f\uff0e BMVC\u3078\u306e\u53c2\u52a0\u306f\u521d\u3081\u3066\u3067\u3057\u305f\u304c\uff0c\u30b3\u30f3\u30d1\u30af\u30c8\u306b\u307e\u3068\u307e\u3063\u305f\u4f1a\u8b70\u3067\u3057\u305f\uff0e\u500b\u4eba\u7684\u306b\u7279\u306b\u6c17\u306b\u306a\u3063\u305f\u767a\u8868\u3092\u5e7e\u3064\u304b\u7d39\u4ecb\u3057\u305f\u3044\u3068\u601d\u3044\u307e\u3059\uff0e \u203b\u79c1\u306e\u524d\u63d0\u77e5\u8b58\u4e0d\u8db3\u3067\uff0c\u7406\u89e3\u304c\u66d6\u6627\u306a\u70b9\u3084\uff0c\u9593\u9055\u3063\u305f\u7406\u89e3\u3092\u7406\u89e3\u3092\u3057\u3066\u3044\u308b\u70b9\u304c\u3042\u308b\u3068\u601d\u3044\u307e\u3059\uff0e\u9593\u9055\u3044\u7b49\u3054\u3056\u3044\u307e\u3057\u305f\u3089\u3054\u6307\u6458\u3044\u305f\u3060\u3051\u307e\u3059\u3068\u5e78\u3044\u3067\u3059\uff0e BMVC BMVC\u306f\uff0c1985\u5e74\u306b\u59cb\u307e\u3063\u305f\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30d3\u30b8\u30e7\u30f3(CV)\u306e\u5b66\u4f1a\u3067\u3059\uff0e \u5b66\u4f1a\u540d\u306b\u306fBritish\u304c\u3064\u3044\u3066\u3044\u307e\u3059\u304c\uff0c\u591a\u304f\u306e\u8ad6\u6587\u304c\u30a4\u30ae\u30ea\u30b9\u56fd\u5916\u304b\u3089\u306e\u3082\u306e\u3067\uff0c\u65e5\u672c\u304b\u3089\u306e\u767a\u8868\u3082\u6570\u4ef6\u3042\u308a\u307e\u3057\u305f\uff0e \u30aa\u30e9\u30fc\u30eb\u767a\u8868\u304c\u30b7\u30f3\u30b0\u30eb\u30c8\u30e9\u30c3\u30af\u306a\u306e\u3067\uff08\u7406\u89e3\u3067\u304d\u308b\u304b\u306f\u5225\u3068\u3057\u3066\uff09\u4e00\u4eba\u3067\u884c\u3063\u3066\u3082\u5168\u90e8\u8074\u8b1b\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\uff0e\u30dd\u30b9\u30bf\u30fc\u3082\u5168\u6587\u3067100\u4ef6\u7a0b\u5ea6\u3067\u3059\uff0e Microsoft Academic\u30e9\u30f3\u30ad\u30f3\u30b0\u3067\u306fCV\u5206\u91ce\u3067\u306f\uff14\u756a\u76ee\u3067\uff0c\u767a\u8868\u6570\u304c\u5c11\u306a\u3044\u5272\u306b\u306f\u5f71\u97ff\u529b\u306e\u3042\u308b\u8ad6\u6587\u304c\u591a\u3044\u3088\u3046\u306b\u611f\u3058\u307e\u3059\uff0e \u4eca\u5e74\u306e\u8ad6\u6587\u306fSupplemental Material\u3092\u542b\u3081\u3066\u516c\u5f0fHP\u3088\u308a\u5168\u90e8\u95b2\u89a7\u3067\u304d\u307e\u3059\uff0e \u6c17\u306b\u306a\u3063\u305f\u767a\u8868\u306e\u4e2d\u304b\u3089\u5e7e\u3064\u304b\u30d4\u30c3\u30af\u30a2\u30c3\u30d7\u3057\u3066\u7d39\u4ecb\u3057\u307e\u3059\uff0e \u79c1\u306e\u8aac\u660e\u3067\u306f\u89e6\u308a\u7a0b\u5ea6\u3057\u304b\u4f1d\u308f\u3089\u306a\u3044\u3068\u304a\u3082\u3044\u307e\u3059\uff0e\u5404\u8ad6\u6587\u6bce\u306b1\u30da\u30fc\u30b8\u306eExtended Abstract\u304c\u516c\u958b\u3055\u308c\u3066\u3044\u307e\u3059\u306e\u3067\uff0c\u305d\u308c\u3092\u773a\u3081\u308b\u306e\u3082\u3044\u3044\u3068\u601d\u3044\u307e\u3059\uff0e [Tutorial] Towards Affordable Self-driving Cars \u30c8\u30ed\u30f3\u30c8\u5927\u306eRaquel Urtasun\u3055\u3093\u306b\u3088\u308b\uff0c\u81ea\u52d5\u904b\u8ee2\u95a2\u9023\u306e\u6280\u8853\u306e\u95a2\u3059\u308b\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\uff0c\u7279\u306b\u5f7c\u5973\u305f\u3061\u306e\u7814\u7a76\u30b0\u30eb\u30fc\u30d7\u306e\u7814\u7a76\u6210\u679c\u306b\u3064\u3044\u3066\u8aac\u660e\u3057\u3066\u3044\u307e\u3057\u305f\uff0e Raquel\u3055\u3093\u306e\u30b0\u30eb\u30fc\u30d7\u306f\uff0cKITTI\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u306a\u3069\u81ea\u52d5\u904b\u8ee2\u306b\u95a2\u3059\u308b\u7814\u7a76\u3092\uff11\u3064\u306e\u67f1\u3068\u3057\u3066\u3044\u3066\uff0cCVPR,NIPS\u306a\u3069\u306e\u30c8\u30c3\u30d7\u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u306b\u6bce\u5e74\u305f\u304f\u3055\u3093\u306e\u8ad6\u6587\u3092\u901a\u3057\u3066\u3044\u307e\u3059\uff0e \u4eca\u56de\u306e\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\u306f\uff0c\u30bf\u30a4\u30c8\u30eb\u306bAffordable\u3068\u3042\u308b\u3088\u3046\u306b\uff0c\u30d9\u30ed\u30c0\u30a4\u30f3\u306a\u3069\u3092\u7528\u3044\u305a\u306b\u30ab\u30e1\u30e9\u306a\u3069\u65e2\u5b58\u306e\u8eca\u8f09\u30bb\u30f3\u30b7\u30f3\u30b0\u30c7\u30d0\u30a4\u30b9\u306b\u8fd1\u3044\u69cb\u6210\u3067\u81ea\u52d5\u904b\u8ee2\u3092\u5b9f\u73fe\u3059\u308b\u305f\u3081\u306e\u7814\u7a76\u306b\u3064\u3044\u3066\u306e\u767a\u8868\u3057\u3066\u3044\u307e\u3057\u305f\uff0e \u8eca\u8f09\u5358\u773c\u30ab\u30e1\u30e9\u753b\u50cf\u306e\u30bb\u30de\u30f3\u30c6\u30a3\u30c3\u30af\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\uff0cOpenStreetMap\uff08OSM\uff09\u306e\u30ea\u30f3\u30af\u30c7\u30fc\u30bf\u3068Visual Oddmetory\u3092\u7528\u3044\u305f\u81ea\u5df1\u4f4d\u7f6e\u63a8\u5b9a\uff0cOSM\u3092\u7528\u3044\u305f\u8eca\u8f09\u30ab\u30e1\u30e9\u753b\u50cf\u3078\u306e\u30e9\u30d9\u30eb\u4ed8\u6280\u8853\u306a\u3069\u304c\u7d39\u4ecb\u3055\u308c\u3066\u3044\u307e\u3057\u305f\uff0e OSM\u3068Visual Oddmetory\u3092\u7528\u3044\u305f\u81ea\u5df1\u4f4d\u7f6e\u63a8\u5b9a\u3067\u306f\uff0c\u6570\u30ad\u30ed\uff58\u6570\u30ad\u30ed\u56db\u65b9\u306e\u5e83\u7bc4\u56f2\u30a8\u30ea\u30a2\u304b\u3089Visual Oddmetory\u306e\u60c5\u5831\u3092OSM\u306e\u30ea\u30f3\u30af\u60c5\u5831\u3068\u30b0\u30e9\u30d5\u3092\u7528\u3044\u3066\u30de\u30c3\u30c1\u30f3\u30b0\u3059\u308b\u3053\u3068\u3067\uff0c\u52b9\u7387\u7684\u306b\u73fe\u5728\u4f4d\u7f6e\u306e\u63a8\u5b9a\u3092\u884c\u3046\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3057\u305f\uff0e\u554f\u984c\u8a2d\u5b9a\u3084\u539f\u7406\u304c\u3068\u3066\u3082\u9762\u767d\u3044\u3068\u601d\u3044\u307e\u3057\u305f\uff0e\u3053\u3053\u306b\u30d3\u30c7\u30aa\u304c\u516c\u958b\u3055\u308c\u3066\u3044\u307e\u3059\uff0e Adding Synchronization and Rolling Shutter in Multi-Camera Bundle Adjustment \u591a\u304f\u306e\u30b3\u30f3\u30b9\u30fc\u30de\u5411\u3051\u30ab\u30e1\u30e9\u306f\uff0cRolling Shutter\u3068\u547c\u3070\u308c\u308b\u753b\u7d20\u3054\u3068\u306b\u8f1d\u5ea6\u5024\u304c\u30b7\u30fc\u30b1\u30f3\u30b7\u30e3\u30eb\u306b\u8aad\u307f\u8fbc\u307e\u308c\u308b\u69cb\u6210\u306b\u306a\u3063\u3066\u3044\u307e\u3059\uff0e \u3064\u307e\u308a\uff0c\u753b\u7d20\u306e\u5de6\u4e0a\u3068\u53f3\u4e0b\u3067\u306f\uff0c\u6642\u9593\u7684\u306b\u7570\u306a\u308b\u30bf\u30a4\u30df\u30f3\u30b0\u306e\u8f1d\u5ea6\u60c5\u5831\u304c\u5f97\u3089\u308c\u308b\u3053\u3068\u306b\u306a\u308a\u307e\u3059\uff0e \u5f7c\u3089\u306f\uff0c\u30de\u30eb\u30c1\u30ab\u30e1\u30e9\u306eBundle Adjustment\u306b\u304a\u3044\u3066\uff0c\u30ab\u30e1\u30e9\u9593\u306eRolling Shutter\u306e\u540c\u671f\u305a\u308c\u3082\u30d1\u30e9\u30e1\u30fc\u30bf\u3068\u3057\u3066\u52a0\u3048\u305f\u5b9a\u5f0f\u5316\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\uff0e GoPro\u30ab\u30e1\u30e9\u3092\uff14\u3064\u7528\u3044\u305f\u5168\u5468\u30ab\u30e1\u30e9\u3092\u5236\u4f5c\u3057\uff0c\u305d\u308c\u3092\u7528\u3044\u305fSLAM\u306e\u7d50\u679c\u3092\u7d39\u4ecb\u3057\u3066\u3044\u307e\u3057\u305f\uff0e\u767a\u8868\u3067\u306fLoop Clousure\u51e6\u7406\u306a\u3057\u3067\u3082\u30c9\u30ea\u30d5\u30c8\u304c\u5c11\u306a\u3044\u7d50\u679c\u304c\u5f97\u3089\u308c\u308b\u3053\u3068\u3092\u30a2\u30d4\u30fc\u30eb\u3057\u3066\u3044\u307e\u3057\u305f\uff0e Event Camera\u3092\u7528\u3044\u305f\u7814\u7a76 \u4eca\u56de\uff0cEvent Camera\u3092\u7528\u3044\u305f\u767a\u8868\u304c\uff13\u4ef6\u3042\u308a\u307e\u3057\u305f\uff0e EMVS: Event-based Multi-View Stereo, Real-Time Intensity-Image Reconstruction for Event Cameras Using Manifold Regularisation, Event-Based Hough Transform in a Spiking Neural Network for Multiple Line Detection and Tracking Using a […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_locale":"en_US","_original_post":"https:\/\/d-itlab.co.jp\/?p=601","_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[1],"tags":[],"kind_blog":[],"member":[27],"keyword_tech":[],"class_list":["post-1393","post","type-post","status-publish","format-standard","hentry","category-uncategorized","member-27","en-US"],"acf":[],"aioseo_notices":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/posts\/1393","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/comments?post=1393"}],"version-history":[{"count":1,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/posts\/1393\/revisions"}],"predecessor-version":[{"id":1394,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/posts\/1393\/revisions\/1394"}],"wp:attachment":[{"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/media?parent=1393"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/categories?post=1393"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/tags?post=1393"},{"taxonomy":"kind_blog","embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/kind_blog?post=1393"},{"taxonomy":"member","embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/member?post=1393"},{"taxonomy":"keyword_tech","embeddable":true,"href":"https:\/\/d-itlab.co.jp\/wp-json\/wp\/v2\/keyword_tech?post=1393"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}