search

nie-lang / Multi-Grid-Deep-Homography

TCSVT2021 - Multi-Grid Deep Homogarphy
85 stars 7 forks source link

video image stitch #8

Open weidi0 opened 2 years ago

weidi0 commented 2 years ago

hello,I have a question that can this learning-based method be used for video stitching?thank you for your advise.

nie-lang commented 2 years ago

hello,I have a question that can this learning-based method be used for video stitching?thank you for your advise.

It can be used for alignment, but it can not be used for stitching due to the Backward Multi-Grid Deformation. Or if the mesh resolution is set to 1*1 (equal to a global homography), it can be used for stitching.

weidi0 commented 2 years ago

Thank you for your reply.I'm doing the research about video stitching.And I have found your several excellent articles about image stitching.Can I talk to you about the video stitching?At the begining,I want to use the traditional methods to finish my task.However,I don't know if these method are out of style because of the development of the learning-method in the image stitching.I have some ideas and doubt to finish the task of video stitching by using the learning based method.First，is it possible that we can use the stitching model，estimated homography，which is gotten from the result that learning method is used for the first pair of imput images，to stitch the follow frames in the video in which we don't consider there are moving foreground objects.Can the homography be a global homography and multi local homographies？Thus can I pay attention to the research about homography estimation for my task?Second，is it impractical to use the learning method which involve the image alignment and image reconstruction to stitch two input images of every frame？I think the reason may be the dataset and the stitching time problem.There must be problem about my idea and can you give me some advise?