Multimodal Distillation for Egocentric Action Recognition

• URL: http://arxiv.org/abs/2307.07483v2
• Date: Tue, 18 Jul 2023 09:26:57 GMT
• Title: Multimodal Distillation for Egocentric Action Recognition
• Authors: Gorjan Radevski, Dusan Grujicic, Marie-Francine Moens, Matthew Blaschko, Tinne Tuytelaars
• Abstract summary: egocentric video understanding involves modelling hand-object interactions. Standard models, e.g. CNNs or Vision Transformers, which receive RGB frames as input perform well. But their performance improves further by employing additional input modalities that provide complementary cues. The goal of this work is to retain the performance of such a multimodal approach, while using only the RGB frames as input at inference time.
• Score: 41.821485757189656
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The focal point of egocentric video understanding is modelling hand-object interactions. Standard models, e.g. CNNs or Vision Transformers, which receive RGB frames as input perform well. However, their performance improves further by employing additional input modalities that provide complementary cues, such as object detections, optical flow, audio, etc. The added complexity of the modality-specific modules, on the other hand, makes these models impractical for deployment. The goal of this work is to retain the performance of such a multimodal approach, while using only the RGB frames as input at inference time. We demonstrate that for egocentric action recognition on the Epic-Kitchens and the Something-Something datasets, students which are taught by multimodal teachers tend to be more accurate and better calibrated than architecturally equivalent models trained on ground truth labels in a unimodal or multimodal fashion. We further adopt a principled multimodal knowledge distillation framework, allowing us to deal with issues which occur when applying multimodal knowledge distillation in a naive manner. Lastly, we demonstrate the achieved reduction in computational complexity, and show that our approach maintains higher performance with the reduction of the number of input views. We release our code at https://github.com/gorjanradevski/multimodal-distillation.

Related papers

• ADEM-VL: Adaptive and Embedded Fusion for Efficient Vision-Language Tuning [38.26304604660713]
  ADEM-VL is an efficient vision-language method that tunes models based on pretrained large language models. Our framework surpasses existing methods by an average accuracy of 0.77% on ScienceQA dataset.
  arXiv  Detail & Related papers  (2024-10-23T11:31:06Z)
• Towards a Generalist and Blind RGB-X Tracker [91.36268768952755]
  We develop a single model tracker that can remain blind to any modality X during inference time. Our training process is extremely simple, integrating multi-label classification loss with a routing function. Our generalist and blind tracker can achieve competitive performance compared to well-established modal-specific models.
  arXiv  Detail & Related papers  (2024-05-28T03:00:58Z)
• Bi-directional Adapter for Multi-modal Tracking [67.01179868400229]
  We propose a novel multi-modal visual prompt tracking model based on a universal bi-directional adapter. We develop a simple but effective light feature adapter to transfer modality-specific information from one modality to another. Our model achieves superior tracking performance in comparison with both the full fine-tuning methods and the prompt learning-based methods.
  arXiv  Detail & Related papers  (2023-12-17T05:27:31Z)
• Dissecting Multimodality in VideoQA Transformer Models by Impairing Modality Fusion [54.33764537135906]
  VideoQA Transformer models demonstrate competitive performance on standard benchmarks. Do these models capture the rich multimodal structures and dynamics from video and text jointly? Are they achieving high scores by exploiting biases and spurious features?
  arXiv  Detail & Related papers  (2023-06-15T06:45:46Z)
• eP-ALM: Efficient Perceptual Augmentation of Language Models [70.47962271121389]
  We propose to direct effort to efficient adaptations of existing models, and propose to augment Language Models with perception. Existing approaches for adapting pretrained models for vision-language tasks still rely on several key components that hinder their efficiency. We show that by freezing more than 99% of total parameters, training only one linear projection layer, and prepending only one trainable token, our approach (dubbed eP-ALM) significantly outperforms other baselines on VQA and Captioning.
  arXiv  Detail & Related papers  (2023-03-20T19:20:34Z)
• Students taught by multimodal teachers are superior action recognizers [41.821485757189656]
  The focal point of egocentric video understanding is modelling hand-object interactions. Standard models -- CNNs, Vision Transformers, etc. -- which receive RGB frames as input perform well, however, their performance improves further by employing additional modalities such as object detections, optical flow, audio, etc. The goal of this work is to retain the performance of such multimodal approaches, while using only the RGB images as input at inference time.
  arXiv  Detail & Related papers  (2022-10-09T19:37:17Z)
• Leveraging Uni-Modal Self-Supervised Learning for Multimodal Audio-Visual Speech Recognition [23.239078852797817]
  We leverage uni-modal self-supervised learning to promote the multimodal audio-visual speech recognition (AVSR) In particular, we first train audio and visual encoders on a large-scale uni-modal dataset, then we integrate components of both encoders into a larger multimodal framework. Our model is experimentally validated on both word-level and sentence-level AVSR tasks.
  arXiv  Detail & Related papers  (2022-02-24T15:12:17Z)
• Mutual Modality Learning for Video Action Classification [74.83718206963579]
  We show how to embed multi-modality into a single model for video action classification. We achieve state-of-the-art results in the Something-Something-v2 benchmark.
  arXiv  Detail & Related papers  (2020-11-04T21:20:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.