AI Paper Atlas

Research map / Video / Video segmentation and tracking

Video Summarization Using Fully Convolutional Sequence Networks

Mrigank Rochan, Linwei Ye, Yang Wang

ECCV 2018 · Video › Video segmentation and tracking · approach: Video semantic and panoptic segmentation · 302 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace bilinear upsampling with learnable deconvolution (ablation)SUM-DeepLab · TVSum (transfer)F-score: 57.4 -> 58.8Improved
replace bilinear upsampling with learnable deconvolution (ablation)SUM-DeepLab · SumMe (transfer)F-score: 45.0 -> 45.1Improved
augment training data with other datasetsSUM-FCN · SumMeF-score: 47.5 -> 51.1Improved
augment training data with other datasetsSUM-FCN · TVSumF-score: 56.8 -> 59.2Improved
sample videos to T=640 frames instead of 320 (ablation)SUM-FCN · SumMe (transfer)F-score: 44.1 -> 45.6No clear effect
sample videos to T=640 frames instead of 320 (ablation)SUM-FCN · TVSum (transfer)F-score: 58.2 -> 57.4No clear effect
sample videos to T=640 frames instead of 320 (ablation)SUM-DeepLab · SumMe (transfer)F-score: 45.0 -> 44.5No clear effect
process variable-length videos without sampling (ablation)SUM-FCN · SumMe (transfer)F-score: 44.1 -> 46.0Improved
process variable-length videos without sampling (ablation)SUM-FCN · TVSum (transfer)F-score: 58.2 -> 56.7No clear effect
use SUM-DeepLab instead of SUM-FCNSUM-FCN · SumMe (standard supervised)F-score: 47.5 -> 48.8Improved
use SUM-DeepLab instead of SUM-FCNSUM-FCN · TVSum (transfer)F-score: 58.2 -> 57.4Hurt
use fully convolutional sequence network instead of LSTM-based dppLSTMdppLSTM · SumMe (standard supervised)F-score: 38.6 -> 47.5Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy