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 Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Few-shot visual recognition refers to recognize novel visual concepts from afew labeled instances. Many few-shot visual recognition methods adopt themetric-based meta-learning paradigm by comparing the query representation withclass representations to predict the category of query instance. However,current metric-based methods generally treat all instances equally andconsequently often obtain biased class representation, considering not allinstances are equally significant when summarizing the instance-levelrepresentations for the class-level representation. For example, some instancesmay contain unrepresentative information, such as too much background andinformation of unrelated concepts, which skew the results. To address the aboveissues, we propose a novel metric-based meta-learning framework termedinstance-adaptive class representation learning network (ICRL-Net) for few-shotvisual recognition. Specifically, we develop an adaptive instance revaluingnetwork with the capability to address the biased representation issue whengenerating the class representation, by learning and assigning adaptive weightsfor different instances according to their relative significance in the supportset of corresponding class. Additionally, we design an improved bilinearinstance representation and incorporate two novel structural losses, i.e.,intra-class instance clustering loss and inter-class representationdistinguishing loss, to further regulate the instance revaluation process andrefine the class representation. We conduct extensive experiments on fourcommonly adopted few-shot benchmarks: miniImageNet, tieredImageNet, CIFAR-FS,and FC100 datasets. The experimental results compared with the state-of-the-artapproaches demonstrate the superiority of our ICRL-Net.

Not All Instances Contribute Equally: Instance-adaptive Class Representation Learning for Few-Shot Visual Recognition