Fork me on GitHub

Semantic segmentation of fashion images using feature pyramid networks

From left to right: the input image, the ground truth segmentation, the predicted segmentation (ResNeXtFPN), the prediction with CRF, and the incorrectly classi- fied pixels (shown in black) for two top scoring test data image predictions from refined Fashionista.

We approach fashion image analysis through semantic segmentation of fashion images, using both textural information and cues from shape and context, where target classes are clothing categories. Our main contributions are state-of-the-art semantic segmentation of fashion images with modest memory and compute requirements.

John Martinsson, Olof Mogren

Second Workshop on Computer Vision for Fashion, Art and Design at ICCV 2019.
PDF Fulltext bibtex.

Olof Mogren, PhD, RISE Research institutes of Sweden