Abstract:With a wide array of multi-modal, multi-protocol, and multi-scale biomedical data available for disease diagnosis and prognosis, there is a need for quantitative tools to combine such varied channels of information, especially imaging and non-imaging data (e.g. spectroscopy, proteomics). The major problem in such quantitative data integration lies in reconciling the large spread in the range of dimensionalities and scales across the different modalities. The primary goal of quantitative data integration is to build combined meta-classifiers; however these efforts are thwarted by challenges in (1) homogeneous representation of the data channels, (2) fusing the attributes to construct an integrated feature vector, and (3) the choice of learning strategy for training the integrated classifier. In this paper, we seek to (a) define the characteristics that guide the 4 independent methods for quantitative data fusion that use the idea of a meta-space for building integrated multi-modal, multi-scale meta-classifiers, and (b) attempt to understand the key components which allowed each method to succeed. These methods include (1) Generalized Embedding Concatenation (GEC), (2) Consensus Embedding (CE), (3) Semi-Supervised Multi-Kernel Graph Embedding (SeSMiK), and (4) Boosted Embedding Combination (BEC). In order to evaluate the optimal scheme for fusing imaging and non-imaging data, we compared these 4 schemes for the problems of combining (a) multi-parametric MRI with spectroscopy for prostate cancer (CaP) diagnosis in vivo, and (b) histological image with proteomic signatures (obtained via mass spectrometry) for predicting prognosis in CaP patients. The kernel combination approach (SeSMiK) marginally outperformed the embedding combination schemes. Additionally, intelligent weighting of the data channels (based on their relative importance) appeared to outperform unweighted strategies. All 4 strategies easily outperformed a naïve decision fusion approach, suggesting that data integration methods will play an important role in the rapidly emerging field of integrated diagnostics and personalized healthcare.

Health diagnostics using multi-attribute classification fusion.

Design Of A Robust Classification Fusion Platform For Structural Health Diagnostics

Fuzzy Decision Fusion System for Fault Classification with Analytic Hierarchy Process Approach

A Novel Multi-Sensor Hybrid Fusion Framework

Fault Diagnosis Based On Fusion Of Multi-Channel Observations By Sensors With Independent Component Analysis And Mutual Information

A hierarchical fusion framework to integrate homogeneous and heterogeneous classifiers for medical decision-making

Intuitionistic Fuzzy Reasoning for Multiple Two-Class Classifiers Fusion

Multimodal feature fusion in deep learning for comprehensive dental condition classification

AliFuse: Aligning and Fusing Multi-modal Medical Data for Computer-Aided Diagnosis

Famcdm: A Fusion Approach of Mcdm Methods to Rank Multiclass Classification Algorithms

A novel multi-classifier information fusion based on Dempster-Shafer theory: application to vibration-based fault detection

MULTI-MODAL DATA FUSION SCHEMES FOR INTEGRATED CLASSIFICATION OF IMAGING AND NON-IMAGING BIOMEDICAL DATA

Multimodal medical image fusion and classification using deep learning techniques

Multi-task fusion for improving mammography screening data classification

Multi-Classifier Fusion Approach Based On Data Clustering For Analog Circuits Fault Diagnosis

Transformer-Based Multi-Modal Data Fusion Method for COPD Classification and Physiological and Biochemical Indicators Identification

MFA-DMFS:new Multi-Classifier Fusion Method and Its Application

Deep Multi-modal Fusion of Image and Non-image Data in Disease Diagnosis and Prognosis: A Review

Hierarchical cognize framework for the multi-fault diagnosis of the interconnected system based on domain knowledge and data fusion

Multimodal Information Fusion for Glaucoma and DR Classification

Multi-class probability SVM fusion using fuzzy integral for fault diagnosis