Win-Fail Action Recognition

TL;DR

This paper introduces the novel task of win-fail action recognition to differentiate successful from failed attempts in videos, supported by a new dataset and analysis of current methods' limitations in achieving true action understanding.

Contribution

It presents the first paired win-fail action dataset across diverse domains and analyzes existing recognition models' performance and challenges in this new task.

Findings

Current methods perform well but leave a large gap for true understanding.

High intra-class variation makes the task challenging yet feasible.

The dataset enables systematic analysis of win-fail action recognition.

Abstract

Current video/action understanding systems have demonstrated impressive performance on large recognition tasks. However, they might be limiting themselves to learning to recognize spatiotemporal patterns, rather than attempting to thoroughly understand the actions. To spur progress in the direction of a truer, deeper understanding of videos, we introduce the task of win-fail action recognition -- differentiating between successful and failed attempts at various activities. We introduce a first of its kind paired win-fail action understanding dataset with samples from the following domains: "General Stunts," "Internet Wins-Fails," "Trick Shots," and "Party Games." Unlike existing action recognition datasets, intra-class variation is high making the task challenging, yet feasible. We systematically analyze the characteristics of the win-fail task/dataset with prototypical action recognition networks and a novel video retrieval task. While current action recognition methods work well on our task/dataset, they still leave a large gap to achieve high performance. We hope to motivate more work towards the true understanding of actions/videos. Dataset will be available from https://github.com/ParitoshParmar/Win-Fail-Action-Recognition.