Biomedical Imaging
1936–1946
X-ray driven imaging and fluorescence microscopy established core tools for visualizing structure and molecular contrast. Early computational methods and electroencephalography advanced functional brain mapping and vascular measurements. Chromosomal imaging under X-ray exposure began documenting cytogenetic phenomena. These developments laid the foundational framework uniting structural, functional, and genetic observation.
Period 1936-1946: X-ray imaging, fluorescence microscopy, and electroencephalography mapped biological structure and function[1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16]. Foundational atlases and localization studies shaped modern cellular imaging, neurodiagnostics, and ophthalmology[12][17][18][6].
1947–1981
Biomedical imaging underwent a major transformation through the merger of computation, physics, and instrumentation to enable noninvasive visualization. Algorithmic and Fourier-based tomographic reconstruction unlocked three-dimensional views from radiographs and cross-sectional data. Meanwhile, magnetic resonance, ultrasound, and electron microscopy expanded functional assessment across cellular and tissue scales. Ultimately, advanced reconstruction, attenuation correction, and radiotracer validation solidified a quantitative tomographic framework that unified brain, chest, and body imaging.
Period 1947-1974: Volumetric imaging advanced through algorithmic reconstructions[1][2][3], microscopy[4][5][6][7][8][9][10][11], and NMR diagnostics[12][13][14][15], staining[16][17], and tracers[12][6][14]. Pioneering scanner systems[1], Fourier mathematics[18], NMR tumor detection[12], and brain atlases[19] established modern tomographic and neuroimaging standards. Period 1975-1981: Computed tomography unified body imaging while emission tomography advanced quantitative brain scans, thallium perfusion, and metabolism mapping[20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37]. Foundational reconstruction, attenuation correction, and FDG validation studies shaped cross-modality quantitation standards and catalyzed decades of functional imaging[38][20][27][39][36].
1982–2009
Functional magnetic resonance imaging and diffusion tensor imaging revolutionized brain mapping and structural pathway analysis. Automated image analysis, multimodal registration, and quantitative morphometry strengthened cross-modality workflows and reproducibility. Targeted molecular imaging integrated fluorescent nanoscale probes and engineered nanoparticles to visualize biological processes in vivo. Compressed sensing and advanced diffusion methods enabled rapid, quantitative acquisitions across magnetic resonance and optical modalities.
Period 1982-1995: Functional MRI, multimodal registration, diffusion and perfusion imaging, automated morphometry, and optical tools mapped brain function and microstructure[1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18]. Breakthroughs in OCT, super-resolution microscopy, BOLD fMRI, diffusion tensor imaging, and two-photon microscopy established modern multiscale neuroimaging and biomarker pipelines[19][20][21][22][23]. Period 1996-2002: Diffusion MRI and automated registration enabled white-matter connectivity mapping and multi-modal workflows alongside robust bias correction[24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39]. Foundational VMD, AFNI, tractography, DTI, and VBM works shaped later connectomics, open pipelines, and structural neuroimaging research[40][29][24][41][42]. Period 2003-2009: Targeted molecular imaging integrated ligand-guided contrast, optical probes, multimodal nanoparticles, and diffusion MRI uncertainty analysis[43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58]. Sparse reconstruction, tractography atlases, cortical parcellation, and Cre reporters catalyzed rapid scans and brain-wide connectivity studies[59][60][61][62].
2010–2024
Biomedical imaging integrated deep learning, radiomics, and multimodal approaches to enable quantitative analysis and image-guided interventions. Convolutional neural networks revolutionized segmentation and diagnostic tasks across diverse medical modalities. High-throughput radiomics features mapped tumor phenotypes and predicted clinical outcomes using noninvasive imaging. End-to-end deep learning systems ultimately automated detection and diagnosis while addressing data scarcity through transfer learning and standardization.
Period 2010-2016: Researchers combined targeted molecular probes, optical imaging depth advances, diffusion neuroimaging metrics, and theranostics to map and treat biological processes[1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. Network architectures, AlexNet, volumetric CNNs, and radiomics transformed biomedical image segmentation, large-scale vision, and precision oncology[18][19][20][21]. Period 2017-2024: Deep learning models automate detection, segmentation, and adaptation across imaging modalities using large datasets[22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]. Foundational architectures and radiomics systems catalyzed subsequent radiology AI research, precision medicine, and clinical trust[22][23][39][40][41].
X-Ray and Fluorescence Imaging Foundations era
Karl Sax[1] conducted research while affiliated with Harvard University[3] and the U.S. National Arboretum[4]. Through CHROMOSOME ABERRATIONS INDUCED BY X-RAYS[7] and TYPES AND FREQUENCIES OF CHROMOSOMAL ABERRATIONS INDUCED BY X-RAYS[8], they advanced cytogenetic observation by documenting chromosomal phenomena under X-ray exposure. Frederic Wood Jones[2] performed investigations at Google (United States)[5] and the University of Manchester[6]. Through The measurement of particle size by the X-ray method[9], they contributed to X-ray driven imaging by establishing foundational methods for quantifying structural dimensions.
Emergence of Computed Tomographic Imaging era
Klaus G. Bensch[10] conducted research while affiliated with Yale University[13] and Stanford University[14]. Their work in CYTOCHEMISTRY AND ELECTRON MICROSCOPY[17] expanded functional assessment across cellular and tissue scales by applying electron microscopy to biomedical imaging. David D. Sabatini[11] pursued investigations at Rockefeller University[15] and New York University[16]. Through CYTOCHEMISTRY AND ELECTRON MICROSCOPY[17], they contributed to the merger of instrumentation and cytology to enable noninvasive cellular visualization. Russell J. Barrnett[12] performed studies at Yale University[13] and Harvard University[3]. They advanced the field's technological framework through CYTOCHEMISTRY AND ELECTRON MICROSCOPY[17] by enhancing high-resolution structural assessment at the microscopic scale.
Functional and Molecular Multimodal Imaging era
Klaus Schulten[18] conducted research while affiliated with the University of Illinois Urbana-Champaign[21] and the University of Illinois System[22]. Through the work published in VMD: Visual molecular dynamics[24], they contributed to targeted molecular imaging by providing essential visualization tools for nanoscale structures and biological processes in vivo. Andrew Dalke[19] performed investigations at the University of Illinois Urbana-Champaign[21] and Urbana University[23]. Their contribution to the era through VMD: Visual molecular dynamics[24] advanced the computational visualization of molecular systems that supports quantitative and multimodal workflows. William Humphrey[20] likewise worked at the University of Illinois Urbana-Champaign[21] and Urbana University[23]. By helping develop VMD: Visual molecular dynamics[24], they advanced the integration of molecular-level visualization into the broader framework of biological imaging and analysis.
Deep Learning and Quantitative Radiomics era
Geoffrey E. Hinton[25], working at the University of Toronto[28] and Google DeepMind (United Kingdom)[29], laid foundational deep learning architectures for biomedical imaging. Through ImageNet Classification with Deep Convolutional Neural Networks[30], this work introduced convolutional neural networks that ultimately revolutionized segmentation and diagnostic tasks across diverse medical modalities. Ilya Sutskever[26], affiliated with Google (United States)[5] and the University of Toronto[28], advanced computational models central to image analysis. Via ImageNet Classification with Deep Convolutional Neural Networks[30], the research contributed to end-to-end deep learning systems that automated detection and diagnosis while addressing data scarcity. Alex Krizhevsky[27], based at the University of Toronto[28] and Google (United States)[5], developed high-capacity neural network models for visual recognition. By publishing ImageNet Classification with Deep Convolutional Neural Networks[30], this effort powered quantitative analysis and image-guided interventions through advanced feature extraction.
X-Ray and Fluorescence Imaging Foundations
Emergence of Computed Tomographic Imaging
Functional and Molecular Multimodal Imaging
Deep Learning and Quantitative Radiomics
Karl Sax and Frederic Wood Jones
Klaus G. Bensch and 2 others
Klaus Schulten and 2 others
Geoffrey E. Hinton and 2 others