Weakly-Supervised Action Segmentation with Iterative Soft Boundary Assignment

TL;DR

This paper introduces a scalable and efficient weakly-supervised action segmentation framework using a novel temporal convolutional network and an iterative training strategy, achieving competitive results on benchmark datasets.

Contribution

It proposes a new temporal convolutional network and an iterative soft boundary assignment strategy for weakly-supervised action segmentation, improving scalability and efficiency.

Findings

Achieves competitive or superior performance on Breakfast and Hollywood Extended datasets.

Demonstrates effectiveness of the proposed TCFPN and ISBA methods.

Outperforms existing methods in terms of efficiency and scalability.

Abstract

In this work, we address the task of weakly-supervised human action segmentation in long, untrimmed videos. Recent methods have relied on expensive learning models, such as Recurrent Neural Networks (RNN) and Hidden Markov Models (HMM). However, these methods suffer from expensive computational cost, thus are unable to be deployed in large scale. To overcome the limitations, the keys to our design are efficiency and scalability. We propose a novel action modeling framework, which consists of a new temporal convolutional network, named Temporal Convolutional Feature Pyramid Network (TCFPN), for predicting frame-wise action labels, and a novel training strategy for weakly-supervised sequence modeling, named Iterative Soft Boundary Assignment (ISBA), to align action sequences and update the network in an iterative fashion. The proposed framework is evaluated on two benchmark datasets, Breakfast and Hollywood Extended, with four different evaluation metrics. Extensive experimental results show that our methods achieve competitive or superior performance to state-of-the-art methods.