.. _paper-adal:

Training Convolutional Neural Networks to Score Pneumonia in Slaughtered Pigs
==============================================================================

    - :fa:`circle-check` `10.3390/ani11113290 <https://doi.org/10.3390/ani11113290>`_
    - :fa:`calendar` 17 November 2021
    - :fa:`scroll` `MDPI's Animals <https://www.mdpi.com/journal/animals>`_
    - :fa:`tags` :bdg-primary:`Semantic Segmentation` :bdg-primary:`Convolutional Neural Networks` :bdg-primary:`Food Safety`

The study introduces an AI system to *recognize* and *quantify* enzootic pneumonia-like lesions in slaughtered pigs, potentially replacing the time-consuming task currently done by veterinarians. Implementing AI in this process could streamline inspection procedures, adhere to European legislation regarding food hygiene, and reduce the risk of microbial contamination by avoiding direct handling of carcasses and organs.

The outcome of the study is a Deep Learning model that can segment the lungs of slaughtered pigs and score the presence of lesions. The model was integrated into `ADAL: Automatic Detection of Abbattoir Lesions <https://www.f4tlab.com/adal>`_ and is currently operational.