HASA: Hybrid Architecture Search with Aggregation Strategy for Echinococcosis Classification and Ovary Segmentation in Ultrasound Images

TL;DR

This paper introduces HASA, a hybrid neural architecture search framework combining pre-trained backbones and searched cells with novel operations, optimized for ultrasound image classification and segmentation, achieving efficient and effective models.

Contribution

HASA integrates NAS with expert knowledge, introducing lightweight operations and a re-aggregation strategy to improve ultrasound image analysis.

Findings

Outperforms existing architectures in ultrasound classification and segmentation

Produces lightweight models with high accuracy and efficiency

Validated on large ultrasound datasets with superior results

Abstract

Different from handcrafted features, deep neural networks can automatically learn task-specific features from data. Due to this data-driven nature, they have achieved remarkable success in various areas. However, manual design and selection of suitable network architectures are time-consuming and require substantial effort of human experts. To address this problem, researchers have proposed neural architecture search (NAS) algorithms which can automatically generate network architectures but suffer from heavy computational cost and instability if searching from scratch. In this paper, we propose a hybrid NAS framework for ultrasound (US) image classification and segmentation. The hybrid framework consists of a pre-trained backbone and several searched cells (i.e., network building blocks), which takes advantage of the strengths of both NAS and the expert knowledge from existing convolutional neural networks. Specifically, two effective and lightweight operations, a mixed depth-wise convolution operator and a squeeze-and-excitation block, are introduced into the candidate operations to enhance the variety and capacity of the searched cells. These two operations not only decrease model parameters but also boost network performance. Moreover, we propose a re-aggregation strategy for the searched cells, aiming to further improve the performance for different vision tasks. We tested our method on two large US image datasets, including a 9-class echinococcosis dataset containing 9566 images for classification and an ovary dataset containing 3204 images for segmentation. Ablation experiments and comparison with other handcrafted or automatically searched architectures demonstrate that our method can generate more powerful and lightweight models for the above US image classification and segmentation tasks.