Login | Register

Adaptive Memory Management in Video Object Segmentation


Adaptive Memory Management in Video Object Segmentation

Pourganjalikhan, Ali (2021) Adaptive Memory Management in Video Object Segmentation. Masters thesis, Concordia University.

[thumbnail of Pourganjalikhan_MSc_S2022.pdf]
Text (application/pdf)
Pourganjalikhan_MSc_S2022.pdf - Accepted Version
Available under License Spectrum Terms of Access.


Matching-based networks have achieved state-of-the-art performance for video object segmentation (VOS) tasks by storing every-k frames in an external memory bank for future inference. Storing the intermediate frames’ predictions provides the network with richer cues for segmenting an object in the current frame. However, the size of the memory bank gradually increases with the length of the video, which slows down inference speed and makes it impractical to handle arbitrary-length videos.
This thesis proposes an adaptive memory bank strategy for matching-based networks for semi-supervised video object segmentation (VOS) that can handle videos of arbitrary length by discarding obsolete features. Features are indexed based on their importance in the segmentation of the objects in previous frames. Based on the index, we discard unimportant features to accommodate new features. We present our experiments on DAVIS 2016, DAVIS 2017, and Youtube-VOS that
demonstrate that our method outperforms state-of-the-art that employ first-and-latest strategy with fixed-sized memory banks and achieves comparable performance to the every-k strategy with increasing-sized memory banks. Furthermore, experiments show that our method increases inference speed by up to 80% over the very-k and 35% over first-and-latest strategies.
We further investigate memory banks’ attention during the training by proposing two regularizations and studying their effects on performance.

Divisions:Concordia University > Gina Cody School of Engineering and Computer Science > Computer Science and Software Engineering
Item Type:Thesis (Masters)
Authors:Pourganjalikhan, Ali
Institution:Concordia University
Degree Name:M. Sc.
Program:Computer Science
Date:26 November 2021
Thesis Supervisor(s):Poullis, Charalambos
ID Code:990044
Deposited By: Ali Pourganjalikhan
Deposited On:16 Jun 2022 15:01
Last Modified:16 Jun 2022 15:01
All items in Spectrum are protected by copyright, with all rights reserved. The use of items is governed by Spectrum's terms of access.

Repository Staff Only: item control page

Downloads per month over past year

Research related to the current document (at the CORE website)
- Research related to the current document (at the CORE website)
Back to top Back to top