Diagnosing Error in Temporal Action Detectors
This work provides a diagnostic tool for researchers in video understanding to identify specific bottlenecks in temporal action detectors, though it is incremental as it builds on existing methods without introducing a new paradigm.
The authors tackled the problem of understanding progress in temporal action localization by introducing a diagnostic tool to analyze detector performance beyond a single metric, finding that key areas for improvement include handling temporal context, instance size robustness, and reducing localization errors.
Despite the recent progress in video understanding and the continuous rate of improvement in temporal action localization throughout the years, it is still unclear how far (or close?) we are to solving the problem. To this end, we introduce a new diagnostic tool to analyze the performance of temporal action detectors in videos and compare different methods beyond a single scalar metric. We exemplify the use of our tool by analyzing the performance of the top rewarded entries in the latest ActivityNet action localization challenge. Our analysis shows that the most impactful areas to work on are: strategies to better handle temporal context around the instances, improving the robustness w.r.t. the instance absolute and relative size, and strategies to reduce the localization errors. Moreover, our experimental analysis finds the lack of agreement among annotator is not a major roadblock to attain progress in the field. Our diagnostic tool is publicly available to keep fueling the minds of other researchers with additional insights about their algorithms.