Mistake Detection
5 papers with code • 0 benchmarks • 2 datasets
Mistakes are natural occurrences in many tasks and an opportunity for an AR assistant to provide help. Identifying such mistakes requires modelling procedural knowledge and retaining long-range sequence information. In its simplest form Mistake Detection aims to classify each coarse action segment into one of the three classes: {“correct”, “mistake”, “correction”}.
Benchmarks
These leaderboards are used to track progress in Mistake Detection
Assembly101
ImageNet-X
Most implemented papers
Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities
Assembly101 is a new procedural activity dataset featuring 4321 videos of people assembling and disassembling 101 "take-apart" toy vehicles.
PREGO: online mistake detection in PRocedural EGOcentric videos
We propose PREGO, the first online one-class classification model for mistake detection in PRocedural EGOcentric videos.
Differentiable Task Graph Learning: Procedural Activity Representation and Online Mistake Detection from Egocentric Videos
fpv-iplab/differentiable-task-graph-learning
•
•
Task graphs learned with our approach are also shown to significantly enhance online mistake detection in procedural egocentric videos, achieving notable gains of +19. 8% and +7. 5% on the Assembly101-O and EPIC-Tent-O datasets.
TI-PREGO: Chain of Thought and In-Context Learning for Online Mistake Detection in PRocedural EGOcentric Videos
aleflabo/prego
•
•
Mistakes are detected as mismatches between the currently recognized action and the action predicted by the anticipation module.
Task Graph Maximum Likelihood Estimation for Procedural Activity Understanding in Egocentric Videos
fpv-iplab/differentiable-task-graph-learning
•
•
We introduce a gradient-based approach for learning task graphs from procedural activities, improving over hand-crafted methods.
