Generative ScatterNet Hybrid Deep Learning (G-SHDL) Network with Structural Priors for Semantic Image Segmentation

TL;DR

This paper introduces a generative hybrid deep learning network with structural priors that achieves efficient and high-performance semantic image segmentation, especially effective with small labeled datasets.

Contribution

It presents a novel G-SHDL architecture that trains rapidly on limited data and optimizes filter numbers for computational efficiency, outperforming existing methods.

Findings

State-of-the-art performance on two datasets

Effective with small labeled datasets

Computationally efficient architecture

Abstract

This paper proposes a generative ScatterNet hybrid deep learning (G-SHDL) network for semantic image segmentation. The proposed generative architecture is able to train rapidly from relatively small labeled datasets using the introduced structural priors. In addition, the number of filters in each layer of the architecture is optimized resulting in a computationally efficient architecture. The G-SHDL network produces state-of-the-art classification performance against unsupervised and semi-supervised learning on two image datasets. Advantages of the G-SHDL network over supervised methods are demonstrated with experiments performed on training datasets of reduced size.