fdarmon / NeuralWarp

Code release of paper Improving neural implicit surfaces geometry with patch warping
234 stars 14 forks source link

Question about the paper #15

Open DongyangHuLi opened 1 year ago

DongyangHuLi commented 1 year ago

Hi, your great work is impressing! But I'm a little confused about the warp loss and Validity masks: 1."our warping-based loss such that every valid patch in the reference image is given the same weight." why is same weight? 2."second, when the reference and source views are on two different sides of the plane defined by xi and the normal ni; third, when a camera center is too close to the plane defined by xi and the normal ni." Why the binary indicator V will be 0. Looking forward to your reply.