stevramos / video_summarization Goto Github PK

Hey Stev,

I am trying to understand you implementation to generate the dataset. In fact, for the TVSum, I compared the change points you made public with the ones from KaiyangZhou's work. It seems like they are very different despite you are using the same features. I could be some differences in the parameters of the KTS algorithm, but I would like to have your insights.

Second, to generate the user summary for the TVSum dataset, you are thresholding the ground truth scores. However, I think we should apply the knapsack algo on the GT scores to get the shots that belong to the summary.

Thanks for answering my concerns.

Hi Stev,

I'll open this issue to continue our discussion from here, because it isn't about the object features anymore, but rather in general for video summarization datasets.

I highly encourage you to read this paper, Video Summarization Using Deep Neural Networks: A Survey. In particular Table II, page 13, you can see that the most used datasets out there are TVSum and SumMe (OVP and YouTube is mainly for augmentation purposes) and I think that's the reason that every repo out there is using only those h5 files. As a latest trend, I see some new works using VTW dataset, but not the two you mentioned (CoSum and VSUMM).

If you want, you can keep this issue open and keep our discussion live about video summarization in general, and not only about datasets.

George

Hello, I am unable to download the relevant data you mentioned through GShell. Google shows that the request to access "project-367116221053" is invalid. Can you upload the relevant preprocessed dataset and. h5 file?

Hi, I need to build cosum dataset use your code, but when I run "generate_dataset.py" on the raw dataset of cosum, it pop the error "KeyError: 'DOWNLOADED'". Would you please kinly help me? Thanks a lot.

What is mapping from video (.mp4) to shot (.txt)? thanks

Hi Stev,

Your work on summarizing videos using different features was very helpful to me. I was inspired a lot by your work and now I have some doubts about the feature extraction process.

What does rate mean in the shapes of features_rgb, features_flow and features_3D?

I hope you can answer this question for me. Thank you so much!