@article{fdi:010090490,
  title = {{I}mproving robustness of industrial object detection by automatic generation of synthetic images from {CAD} models},
  author = {{B}allhausen {S}ampaio, {I}.{G}. and {V}iterbo, {J}. and {G}uerin, {J}oris},
  editor = {},
  language = {{ENG}},
  abstract = {{O}bject detection ({OD}) is used for visual quality control in factories. {I}mages that compose training datasets are often collected directly from the production line and labeled with bounding boxes manually. {S}uch data represent well the inference context but might lack diversity,  implying a risk of overfitting. {T}o address this issue, we propose a dataset construction method based on an automated pipeline, which receives a {CAD} model of an object and returns a set of realistic synthetic labeled images (code publicly available). {O}ur approach can be easily used by non-expert users and is relevant for industrial applications, where {CAD} models are widely available. {W}e performed experiments to compare the use of datasets obtained by the two different ways  - collecting and labeling real images or applying the proposed automated pipeline -  in the classification of five different industrial parts. {T}o ensure that both approaches can be used without deep learning expertise, all training parameters were kept fixed during these experiments. {I}n our results, both methods were successful for some objects but failed for others. {H}owever, we have shown that the combined use of real and synthetic images led to better results. {T}his finding has the potential to make industrial {OD} models more robust to poor data collection and labeling errors, without increasing the difficulty of the training process.},
  keywords = {},
  booktitle = {},
  journal = {{C}omputational {I}ntelligence},
  volume = {39},
  numero = {3},
  pages = {415--432},
  ISSN = {0824-7935},
  year = {2023},
  DOI = {10.1111/coin.12572},
  URL = {https://www.documentation.ird.fr/hor/fdi:010090490},
}