English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Zero-shot sketch-based image retrieval typically asks for a trained model tobe applied as is to unseen categories. In this paper, we question to argue thatthis setup by definition is not compatible with the inherent abstract andsubjective nature of sketches, i.e., the model might transfer well to newcategories, but will not understand sketches existing in different test-timedistribution as a result. We thus extend ZS-SBIR asking it to transfer to bothcategories and sketch distributions. Our key contribution is a test-timetraining paradigm that can adapt using just one sketch. Since there is nopaired photo, we make use of a sketch raster-vector reconstruction module as aself-supervised auxiliary task. To maintain the fidelity of the trainedcross-modal joint embedding during test-time update, we design a novelmeta-learning based training paradigm to learn a separation between modelupdates incurred by this auxiliary task from those off the primary objective ofdiscriminative learning. Extensive experiments show our model to outperformstate of-the-arts, thanks to the proposed test-time adaption that not onlytransfers to new categories but also accommodates to new sketching styles.

Sketch3T: Test-Time Training for Zero-Shot SBIR