CNN filter sizes, effects, limitations, and challenges: An exploratory study

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Exploratory study reveals opportunities for large filters in cnn architectures, suggesting improved performance outcomes.

MAIN FINDING: Large filter sizes can enhance the performance of cnn models contrary to common beliefs favoring small filters.
KEY EVIDENCE: Current analyses show a bias towards small (3x3) filters, yet large filters exhibit potential for better results.
APPROACH: The study reviews various cnn architectures and emphasizes the limitations and challenges associated with filter sizes.
SIGNIFICANCE: Understanding large filters can help researchers optimize cnn models as computational power and image sizes improve.

Aboukhair et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

View Full Paper

Ask AI

Bookmark

Authors

Exploratory study reveals opportunities for large filters in cnn architectures, suggesting improved performance outcomes.

MAIN FINDING: Large filter sizes can enhance the performance of cnn models contrary to common beliefs favoring small filters.
KEY EVIDENCE: Current analyses show a bias towards small (3x3) filters, yet large filters exhibit potential for better results.
APPROACH: The study reviews various cnn architectures and emphasizes the limitations and challenges associated with filter sizes.
SIGNIFICANCE: Understanding large filters can help researchers optimize cnn models as computational power and image sizes improve.

Aboukhair et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark