Abstract:Learning better representations is essential in medical image analysis for computer-aided diagnosis. However, learning discriminative semantic features is a major challenge due to the lack of large-scale well-annotated datasets. Thus, how can we learn a well-structured categorizable embedding space in limited-scale and unlabeled datasets? In this paper, we proposed a novel clustering-guided twin-contrastive learning framework (CTCL) that learns the discriminative representations of probe-based confocal laser endomicroscopy (pCLE) images for gastrointestinal (GI) tumor classification. Compared with traditional contrastive learning, in which only two randomly augmented views of the same instance are considered, the proposed CTCL aligns more semantically related and class-consistent samples by clustering, which improved intra-class tightness and inter-class variability to produce more informative representations. Furthermore, based on the inherent properties of CLE (geometric invariance and intrinsic noise), we proposed to regard CLE images with any angle rotation and CLE images with different noises as the same instance, respectively, for increased variability and diversity of samples. By optimizing CTCL in an end-to-end expectation-maximization framework, comprehensive experimental results demonstrated that CTCL-based visual representations achieved competitive performance on each downstream task as well as more robustness and transferability compared with existing state-of-the-art SSL and supervised methods. Notably, CTCL achieved 75.60%/78.45% and 64.12%/77.37% top-1 accuracy on the linear evaluation protocol and few-shot classification downstream tasks, respectively, which outperformed the previous best results by 1.27%/1.63% and 0.5%/3%, respectively. The proposed method holds great potential to assist pathologists in achieving an automated, fast, and high-precision diagnosis of GI tumors and accurately determining different stages of tumor development based on CLE images.

Improving the Classification Performance of Esophageal Disease on Small Dataset by Semi-supervised Efficient Contrastive Learning

Supervised Contrastive Learning for Gastrointestinal Lesions Classification in Endoscopic Images

Automatic classification of esophageal disease in gastroscopic images using an efficient channel attention deep dense convolutional neural network

Multi-scale Contrastive Learning for Gastroenteroscopy Classification.

Clustering-Guided Twin Contrastive Learning for Endomicroscopy Image Classification

CLECG: A Novel Contrastive Learning Framework for Electrocardiogram Arrhythmia Classification

Esophageal Squamous Cell Carcinoma Recognition Based on Lightweight Residual Networks with an Attention Mechanism

Improving image classification of gastrointestinal endoscopy using curriculum self-supervised learning

An improved contrastive learning network for semi-supervised multi-structure segmentation in echocardiography

A Simple Framework for Depth-Augmented Contrastive Learning for Endoscopic Image Classification

SSL-CPCD: Self-supervised learning with composite pretext-class discrimination for improved generalisability in endoscopic image analysis

Self-Ensembling Contrastive Learning for Semi-Supervised Medical Image Segmentation

Combating Medical Label Noise via Robust Semi-supervised Contrastive Learning

Deep learning Instance Segmentation on Esophageal Squamous Cell Carcinoma detection

Cervical Cytology Classification with Coarse Labels Based on Two-Stage Weakly Supervised Contrastive Learning Framework.

A Clinical-oriented Multi-level Contrastive Learning Method for Disease Diagnosis in Low-quality Medical Images

Self- and Semi-supervised Learning for Gastroscopic Lesion Detection

Local-global multiple perception based deep multi-modality learning for sub-type of esophageal cancer classification

Multi-Class Gastroesophageal Reflux Disease Classification System Using Deep Learning Techniques

Semi-Supervised Medical Image Classification Combined with Unsupervised Deep Clustering

Self-Supervised Cross-Level Consistency Learning for Fundus Image Classification