Main content

Home

Menu

Loading wiki pages...

View
Wiki Version:
## Crêpe Dataset ![Structured cooking activities][1] ### Summary The Crêpe Dataset provides 6 different types of structured cooking activity videos in 1920x1080 resolution. Each cooking activity is represented as a sequence of different action components. Notable features of this dataset includes: - Structured activities as a sequence of component actions. - Multiple activities running in parallel. - Inclusion of distractors that are not relevant to defined activities. - Every frame is annotated with bounding boxes, agent types, agent occlusion and action labels. We provide the following human-labeled annotations: - Bounding box of every person - Person type (action performer or distractor) - Occlusion against another person - Action label - Activity label Here's a sample processed video superimposed with bounding boxes and action recognition likelihoods: @[youtube](https://youtu.be/wqp5JGANh18) ### Actions 1. cut 2. flip 3. fold 4. grate 5. pour 6. spread 7. sprinkle 8. stir 9. transfer ### Activities | 1. Lemon and sugar | 2. Nutella and banana with chocholate | 3. Cheese and ham | |--------------------------------|---------------------------------------|----------------------------------------| | stir | stir | stir | | pour | pour | pour | | spread | spread | spread | | flip | cut | cut | | pour | flip | grate | | | transfer | flip | | | grate | transfer | | | fold | fold | | | | | | **4. Cheese and ham with parsley** | **5. Goat cheese and spinach** | **6. Goat cheese and spinach with nutmeg** | | stir | stir | stir | | pour | pour | pour | | spread | spread | spread | | cut | cut | cut | | grate | flip | flip | | cut | transfer | transfer | | flip | fold | sprinkle | | transfer | | fold | | sprinkle | | | | fold | | | Please see **Annotation** folder for complete information. ### Citation If you publish using our data set, we would appreciate if you cite: K. Lee, D. Ognibene, H. J. Chang, T. K. Kim and Y. Demiris, "STARE: Spatio-Temporal Attention Relocation for Multiple Structured Activities Detection," in IEEE Transactions on Image Processing, vol. 24, no. 12, pp. 5916-5927, Dec. 2015. For questions and comments, please contact Kyuhwa Lee: lee.kyuh_at_gmail.com ---------- The Crêpe dataset creation was funded by EPSRC Network on Vision and Language, under grant scheme Pump-Priming V&L Research 2013-1. [1]: https://mfr.osf.io/export?url=https://osf.io/z6yat/?action=download%26direct%26mode=render&initialWidth=565&childId=mfrIframe&format=1200x1200.jpeg