From 72082bd96bd13de76f1a4407c0b93577e843434a Mon Sep 17 00:00:00 2001 From: Assembly101 <3dassembly101@gmail.com> Date: Thu, 3 Aug 2023 16:56:22 +0800 Subject: [PATCH] Update README.md --- README.md | 66 +++++++++++++++++++++++++++++++------------------------ 1 file changed, 37 insertions(+), 29 deletions(-) diff --git a/README.md b/README.md index 9d0a953..b2adde2 100644 --- a/README.md +++ b/README.md @@ -1,21 +1,43 @@ -# mistake-detection +# assembly101-mistake-detection +This repository contains annotations for the Mistake Detection benchmark of Assembly101. Please refer to [Every Mistake Counts in Assembly](https://arxiv.org/abs/2307.16453) for further details regarding annotations and baselines. -## Annotation -Each CSV file corresponds to an assembly sequence, and within each CSV file are the annotations. -- There are 328 sequences in total. -- Each annotation contains the - - **start** and **end** timestamps, followed by the - - **verb** (attach, detach) and the - - two working objects (**this, that**), the - - mistake **label** and the - - **remark** are then appended. +If you find these annotations helpful and use them in your research, kindly cite the following: +``` +@article{sener2022assembly101, + title = {Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities}, + author = {F. Sener and D. Chatterjee and D. Shelepov and K. He and D. Singhania and R. Wang and A. Yao}, + journal = {CVPR 2022}, +} + +@article{ding2023every, + title={Every Mistake Counts in Assembly}, + author={Ding, Guodong and Sener, Fadime and Ma, Shugao and Yao, Angela}, + journal={arXiv preprint arXiv:2307.16453}, + year={2023} +} +``` + +## Annotations +Each .csv file provides the mistake annotation for an assembly sequence. +- There are 328 sequences in total. +- Each annotation contains: + - `start` and `end` timestamps + - `verb` (attach, detach) + - two working objects (`this`, `that`) + - mistake `label` + - `remark` (if any) -### Examplar +## File structure -For example: -**nusar-2021_action_both_9033-c02a_9033_user_id_2021-02-04_140532.csv** +For the assembly sequence `nusar-2021_action_both_9033-c02a_9033_user_id_2021-02-04_140532.csv` -- **9033** is the actor while **c02d** is the toy id. +- `9033` is the actor while `c02a` is the toy id. +- `wrong order` : this action is an ordering mistake +- `previous one is mistake` : this action is also an ordering mistake but is caused by the preceding ordering mistakes in the context +- `shouldn't have happened` : this detach action is unnecessary +- `wrong position` : the two parts are not attached at their correct position + +
| start | end | verb | this | that | label | remark | |------------|----------|----------|---------------|---------------|------------|-------------------------| @@ -49,23 +71,9 @@ For example: | 24401 | 25738 | attach | bumper | body | correct | | | 25738 | 27434 | attach | wheel | chassis | correct | | -- **wrong order**: meaning the action is a mistake in ordering -- **previous one is mistake**: meaning this action is also an ordering mistake but caused by preceding ordering mistakes in context -- **shouldn't have happened**: meaning this action is unnecessary detach of parts -- **wrong position**: meaning the two parts are not attached in the right position - -There is no *correction* for the *shouldn't have happened* type of mistake. +**Note**: There is no *correction* for the *shouldn't have happened* type of mistake. -If you find the annotations helpful and use them in your research, kindly cite the following: -``` -@article{ding2023every, - title={Every Mistake Counts in Assembly}, - author={Ding, Guodong and Sener, Fadime and Ma, Shugao and Yao, Angela}, - journal={arXiv preprint arXiv:2307.16453}, - year={2023} -} -``` ## License Assembly101 is licensed by us under the Creative Commons Attribution-NonCommerial 4.0 International License, found [here](https://creativecommons.org/licenses/by-nc/4.0/). The terms are :