Relación entre convolución en matemáticas y CNN

He leído la explicación de la convolución y la entiendo hasta cierto punto. ¿Alguien puede ayudarme a entender cómo esta operación se relaciona con la convolución en redes neuronales convolucionales? ¿Es el filtro como una función gque aplica

machine-learning neural-network deep-learning cnn convolution machine-learning ensemble-modeling machine-learning classification data-mining clustering machine-learning feature-selection convnet pandas graphs ipython machine-learning apache-spark multiclass-classification naive-bayes-classifier multilabel-classification machine-learning data-mining dataset data-cleaning data machine-learning data-mining statistics correlation machine-learning data-mining dataset data-cleaning data beginner career python r visualization machine-learning data-mining nlp stanford-nlp dataset linear-regression time-series correlation anomaly-detection ensemble-modeling data-mining machine-learning python data-mining recommender-system machine-learning cross-validation model-selection scoring prediction sequential-pattern-mining categorical-data python tensorflow image-recognition statistics machine-learning data-mining predictive-modeling data-cleaning preprocessing classification deep-learning tensorflow machine-learning algorithms data keras categorical-data reference-request loss-function classification logistic-regression apache-spark prediction naive-bayes-classifier beginner nlp word2vec vector-space-models scikit-learn decision-trees data programming