MED-VT++: Unifying Multimodal Learning with a Multiscale Encoder-Decoder Video Transformer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Karim, Rezaul, Zhao, He, Wildes, Richard P., Siam, Mennatullah
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910605906018304
author Karim, Rezaul
Zhao, He
Wildes, Richard P.
Siam, Mennatullah
author_facet Karim, Rezaul
Zhao, He
Wildes, Richard P.
Siam, Mennatullah
contents In this paper, we present an end-to-end trainable unified multiscale encoder-decoder transformer that is focused on dense prediction tasks in video. The presented Multiscale Encoder-Decoder Video Transformer (MED-VT) uses multiscale representation throughout and employs an optional input beyond video (e.g., audio), when available, for multimodal processing (MED-VT++). Multiscale representation at both encoder and decoder yields three key benefits: (i) implicit extraction of spatiotemporal features at different levels of abstraction for capturing dynamics without reliance on input optical flow, (ii) temporal consistency at encoding and (iii) coarse-to-fine detection for high-level (e.g., object) semantics to guide precise localization at decoding. Moreover, we present a transductive learning scheme through many-to-many label propagation to provide temporally consistent video predictions. We showcase MED-VT/MED-VT++ on three unimodal video segmentation tasks (Automatic Video Object Segmentation (AVOS), actor-action segmentation and Video Semantic Segmentation (VSS)) as well as a multimodal segmentation task (Audio-Visual Segmentation (AVS)). Results show that the proposed architecture outperforms alternative state-of-the-art approaches on multiple benchmarks using only video (and optional audio) as input, without reliance on optical flow. Finally, to document details of the model's internal learned representations, we present a detailed interpretability study, encompassing both quantitative and qualitative analyses.
format Preprint
id arxiv_https___arxiv_org_abs_2304_05930
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle MED-VT++: Unifying Multimodal Learning with a Multiscale Encoder-Decoder Video Transformer
Karim, Rezaul
Zhao, He
Wildes, Richard P.
Siam, Mennatullah
Computer Vision and Pattern Recognition
In this paper, we present an end-to-end trainable unified multiscale encoder-decoder transformer that is focused on dense prediction tasks in video. The presented Multiscale Encoder-Decoder Video Transformer (MED-VT) uses multiscale representation throughout and employs an optional input beyond video (e.g., audio), when available, for multimodal processing (MED-VT++). Multiscale representation at both encoder and decoder yields three key benefits: (i) implicit extraction of spatiotemporal features at different levels of abstraction for capturing dynamics without reliance on input optical flow, (ii) temporal consistency at encoding and (iii) coarse-to-fine detection for high-level (e.g., object) semantics to guide precise localization at decoding. Moreover, we present a transductive learning scheme through many-to-many label propagation to provide temporally consistent video predictions. We showcase MED-VT/MED-VT++ on three unimodal video segmentation tasks (Automatic Video Object Segmentation (AVOS), actor-action segmentation and Video Semantic Segmentation (VSS)) as well as a multimodal segmentation task (Audio-Visual Segmentation (AVS)). Results show that the proposed architecture outperforms alternative state-of-the-art approaches on multiple benchmarks using only video (and optional audio) as input, without reliance on optical flow. Finally, to document details of the model's internal learned representations, we present a detailed interpretability study, encompassing both quantitative and qualitative analyses.
title MED-VT++: Unifying Multimodal Learning with a Multiscale Encoder-Decoder Video Transformer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2304.05930