CVMar 21, 2019

PProCRC: Probabilistic Collaboration of Image Patches

arXiv:1903.09123v3
Originality Incremental advance
AI Analysis

This work addresses fine-grained species recognition, offering an incremental improvement over existing collaborative representation methods.

The authors tackled the problem of collaborative representation for image patches by proposing a conditional probabilistic framework that integrates background compensation and outlier suppression, eliminating the need for pre-processing. The method outperformed earlier CRC formulations and state-of-the-art probabilistic methods on three fine-grained species recognition datasets, achieving improved results with CNN backbones.

We present a conditional probabilistic framework for collaborative representation of image patches. It incorporates background compensation and outlier patch suppression into the main formulation itself, thus doing away with the need for pre-processing steps to handle the same. A closed form non-iterative solution of the cost function is derived. The proposed method (PProCRC) outperforms earlier CRC formulations: patch based (PCRC, GP-CRC) as well as the state-of-the-art probabilistic (ProCRC and EProCRC) on three fine-grained species recognition datasets (Oxford Flowers, Oxford-IIIT Pets and CUB Birds) using two CNN backbones (Vgg-19 and ResNet-50).

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes