search

ok1zjf / VASNet

PyTorch implementation of the ACCV 2018-AIU2018 paper Video Summarization with Attention
MIT License
180 stars 48 forks source link

input to KTS #5

Open hungbie opened 4 years ago

hungbie commented 4 years ago

Hi,

For change point detection. What should i input to KTS? Flatten image as HxW dimension input or using some feature extraction methods so that the image become some N dimension input? What is being used in this paper to preprocess the image/frame?

18pxrsun commented 4 years ago

Have you knew the input to KTS ?If you have,can you told it to me ?

hungbie commented 4 years ago

You need to process the frame from (h,w) to (d) where d is the number of feature. Basically your video will be from (time,h,w) to (time,d). For that the original KTS use Fisher vector. I used Googlenet instead.

electroncastle commented 4 years ago

Yes, that's right, use the GoogLeNet features as the KTS input.