Object-Centric Cropping for Visual Few-Shot Classification
This addresses performance issues in few-shot learning for computer vision applications, but it is incremental as it builds on existing localization techniques.
The paper tackles the problem of image ambiguities from multiple objects or complex backgrounds degrading performance in Few-Shot Image Classification, showing that using object localization information significantly improves classification across benchmarks, with much of the gain achievable via the Segment Anything Model with minimal annotation or unsupervised methods.
In the domain of Few-Shot Image Classification, operating with as little as one example per class, the presence of image ambiguities stemming from multiple objects or complex backgrounds can significantly deteriorate performance. Our research demonstrates that incorporating additional information about the local positioning of an object within its image markedly enhances classification across established benchmarks. More importantly, we show that a significant fraction of the improvement can be achieved through the use of the Segment Anything Model, requiring only a pixel of the object of interest to be pointed out, or by employing fully unsupervised foreground object extraction methods.