Graph-Based Feature Learning from Image Markers

Abstract

Deep learning methods have achieved impressive results for object detection, but they usually require powerful GPUs and large annotated datasets. In contrast, there is a lack of explainable networks in the literature. For instance, Feature Learning from Image Markers (FLIM) is a feature extraction strategy for lightweight CNNs without backpropagation that requires only a few training images. In this work, we extend FLIM for general image graph modeling, allowing it for a non-strict kernel shape and taking advantage of the adjacency relation between nodes to extract feature vectors based on neighbors’ features. To produce saliency maps by combining learned features, we proposed a User-Guided Decoder (UGD) that does not require training and is suitable for any FLIM-based strategy. Our results indicate that the proposed Graph-based FLIM, named GFLIM, not only outperforms FLIM but also produces competitive detections with deep models, even having an architecture thousands of times smaller in the number of parameters. Our code is publicly available at https://github.com/IMScience-PPGINF-PucMinas/GFLIM.

The authors thank the Pontifícia Universidade Católica de Minas Gerais – PUC-Minas, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES – (Grant COFECUB 88887.191730/2018-00, Grant PROAP 88887.842889/2023-00 – PUC/MG and Finance Code 001), the Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq (Grants 303808/2018-7, 407242/2021-0, 306573/2022-9) and Fundação de Apoio à Pesquisa do Estado de Minas Gerais – FAPEMIG (Grant APQ-01079-23), PUC Minas and INRIA under the project Learning on graph-based hierarchical methods for image and multimedia data.