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/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

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

Zendy Tools

Zendy Open

Automatic Pill Recognition (APR) systems are crucial for enhancing hospitalefficiency, assisting visually impaired individuals, and preventingcross-infection. However, most existing deep learning-based pill recognitionsystems can only perform classification on classes with sufficient trainingdata. In practice, the high cost of data annotation and the continuous increasein new pill classes necessitate the development of a few-shot class-incrementalpill recognition system. This paper introduces the first few-shotclass-incremental pill recognition framework, named Discriminative andBidirectional Compatible Few-Shot Class-Incremental Learning (DBC-FSCIL). Itencompasses forward-compatible and backward-compatible learning components. Inforward-compatible learning, we propose an innovative virtual class synthesisstrategy and a Center-Triplet (CT) loss to enhance discriminative featurelearning. These virtual classes serve as placeholders in the feature space forfuture class updates, providing diverse semantic knowledge for model training.For backward-compatible learning, we develop a strategy to synthesize reliablepseudo-features of old classes using uncertainty quantification, facilitatingData Replay (DR) and Knowledge Distillation (KD). This approach allows for theflexible synthesis of features and effectively reduces additional storagerequirements for samples and models. Additionally, we construct a new pillimage dataset for FSCIL and assess various mainstream FSCIL methods,establishing new benchmarks. Our experimental results demonstrate that ourframework surpasses existing State-of-the-art (SOTA) methods. The code isavailable at https://github.com/zhang-jinghua/DBC-FSCIL.

A Forward and Backward Compatible Framework for Few-shot Class-incremental Pill Recognition