ModelEfficientNet B5
EfficientNet-B5
A mid-to-large-scale variant of Google's EfficientNet convolutional neural network family. It uses compound scaling to simultaneously adjust network depth, width, and input resolution, making it well suited for image tasks that require capturing fine-grained texture information.
Timeline (last 90 days)
Oct 5
The hair shaft counting task is handled by EfficientNet-B5 with a support-map channel, achieving a counting accuracy of 87.0% aligned with expert bounding boxes and a macro-averaged F1 score of 0.85.
Unverified50%