Search

Your search keyword '"Freudiger, Christian W."' showing total 133 results
Start Over Author "Freudiger, Christian W."
133 results on '"Freudiger, Christian W."'

1. Localization of protoporphyrin IX during glioma-resection surgery via paired stimulated Raman histology and fluorescence microscopy

Author

Nasir-Moin, Mustafa, Wadiura, Lisa Irina, Sacalean, Vlad, Juros, Devin, Movahed-Ezazi, Misha, Lock, Emily K., Smith, Andrew, Lee, Matthew, Weiss, Hannah, Müther, Michael, Alber, Daniel, Ratna, Sujay, Fang, Camila, Suero-Molina, Eric, Hellwig, Sönke, Stummer, Walter, Rössler, Karl, Hainfellner, Johannes A., Widhalm, Georg, Kiesel, Barbara, Reichert, David, Mischkulnig, Mario, Jain, Rajan, Straehle, Jakob, Neidert, Nicolas, Schnell, Oliver, Beck, Jürgen, Trautman, Jay, Pastore, Steve, Pacione, Donato, Placantonakis, Dimitris, Oermann, Eric Karl, Golfinos, John G., Hollon, Todd C., Snuderl, Matija, Freudiger, Christian W., Heiland, Dieter Henrik, and Orringer, Daniel A.

Published
2024
Full Text
View/download PDF

2. Hierarchical discriminative learning improves visual representations of biomedical microscopy

Author

Jiang, Cheng, Hou, Xinhai, Kondepudi, Akhil, Chowdury, Asadur, Freudiger, Christian W., Orringer, Daniel A., Lee, Honglak, and Hollon, Todd C.

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

Learning high-quality, self-supervised, visual representations is essential to advance the role of computer vision in biomedical microscopy and clinical medicine. Previous work has focused on self-supervised representation learning (SSL) methods developed for instance discrimination and applied them directly to image patches, or fields-of-view, sampled from gigapixel whole-slide images (WSIs) used for cancer diagnosis. However, this strategy is limited because it (1) assumes patches from the same patient are independent, (2) neglects the patient-slide-patch hierarchy of clinical biomedical microscopy, and (3) requires strong data augmentations that can degrade downstream performance. Importantly, sampled patches from WSIs of a patient's tumor are a diverse set of image examples that capture the same underlying cancer diagnosis. This motivated HiDisc, a data-driven method that leverages the inherent patient-slide-patch hierarchy of clinical biomedical microscopy to define a hierarchical discriminative learning task that implicitly learns features of the underlying diagnosis. HiDisc uses a self-supervised contrastive learning framework in which positive patch pairs are defined based on a common ancestry in the data hierarchy, and a unified patch, slide, and patient discriminative learning objective is used for visual SSL. We benchmark HiDisc visual representations on two vision tasks using two biomedical microscopy datasets, and demonstrate that (1) HiDisc pretraining outperforms current state-of-the-art self-supervised pretraining methods for cancer diagnosis and genetic mutation prediction, and (2) HiDisc learns high-quality visual representations using natural patch diversity without strong data augmentations., Comment: CVPR 2023. Project page: https://hidisc.mlins.org

Published
2023

3. OpenSRH: optimizing brain tumor surgery using intraoperative stimulated Raman histology

Author

Jiang, Cheng, Chowdury, Asadur, Hou, Xinhai, Kondepudi, Akhil, Freudiger, Christian W., Conway, Kyle, Camelo-Piragua, Sandra, Orringer, Daniel A., Lee, Honglak, and Hollon, Todd C.

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning
Abstract

Accurate intraoperative diagnosis is essential for providing safe and effective care during brain tumor surgery. Our standard-of-care diagnostic methods are time, resource, and labor intensive, which restricts access to optimal surgical treatments. To address these limitations, we propose an alternative workflow that combines stimulated Raman histology (SRH), a rapid optical imaging method, with deep learning-based automated interpretation of SRH images for intraoperative brain tumor diagnosis and real-time surgical decision support. Here, we present OpenSRH, the first public dataset of clinical SRH images from 300+ brain tumors patients and 1300+ unique whole slide optical images. OpenSRH contains data from the most common brain tumors diagnoses, full pathologic annotations, whole slide tumor segmentations, raw and processed optical imaging data for end-to-end model development and validation. We provide a framework for patch-based whole slide SRH classification and inference using weak (i.e. patient-level) diagnostic labels. Finally, we benchmark two computer vision tasks: multiclass histologic brain tumor classification and patch-based contrastive representation learning. We hope OpenSRH will facilitate the clinical translation of rapid optical imaging and real-time ML-based surgical decision support in order to improve the access, safety, and efficacy of cancer surgery in the era of precision medicine. Dataset access, code, and benchmarks are available at opensrh.mlins.org., Comment: Neural Information Processing Systems (NeurIPS) 2022 Datasets and Benchmarks Track

Published
2022

4. Near real-time intraoperative brain tumor diagnosis using stimulated Raman histology and deep neural networks.

Author

Hollon, Todd C, Pandian, Balaji, Adapa, Arjun R, Urias, Esteban, Save, Akshay V, Khalsa, Siri Sahib S, Eichberg, Daniel G, D'Amico, Randy S, Farooq, Zia U, Lewis, Spencer, Petridis, Petros D, Marie, Tamara, Shah, Ashish H, Garton, Hugh JL, Maher, Cormac O, Heth, Jason A, McKean, Erin L, Sullivan, Stephen E, Hervey-Jumper, Shawn L, Patil, Parag G, Thompson, B Gregory, Sagher, Oren, McKhann, Guy M, Komotar, Ricardo J, Ivan, Michael E, Snuderl, Matija, Otten, Marc L, Johnson, Timothy D, Sisti, Michael B, Bruce, Jeffrey N, Muraszko, Karin M, Trautman, Jay, Freudiger, Christian W, Canoll, Peter, Lee, Honglak, Camelo-Piragua, Sandra, and Orringer, Daniel A

Subjects
Humans, Brain Neoplasms, Monitoring, Intraoperative, Spectrum Analysis, Raman, Probability, Algorithms, Computer Systems, Image Processing, Computer-Assisted, Clinical Trials as Topic, Deep Learning, Neural Networks, Computer, Monitoring, Intraoperative, Spectrum Analysis, Raman, Image Processing, Computer-Assisted, Neural Networks, Computer, Medical and Health Sciences, Immunology
Abstract

Intraoperative diagnosis is essential for providing safe and effective care during cancer surgery1. The existing workflow for intraoperative diagnosis based on hematoxylin and eosin staining of processed tissue is time, resource and labor intensive2,3. Moreover, interpretation of intraoperative histologic images is dependent on a contracting, unevenly distributed, pathology workforce4. In the present study, we report a parallel workflow that combines stimulated Raman histology (SRH)5-7, a label-free optical imaging method and deep convolutional neural networks (CNNs) to predict diagnosis at the bedside in near real-time in an automated fashion. Specifically, our CNNs, trained on over 2.5 million SRH images, predict brain tumor diagnosis in the operating room in under 150 s, an order of magnitude faster than conventional techniques (for example, 20-30 min)2. In a multicenter, prospective clinical trial (n = 278), we demonstrated that CNN-based diagnosis of SRH images was noninferior to pathologist-based interpretation of conventional histologic images (overall accuracy, 94.6% versus 93.9%). Our CNNs learned a hierarchy of recognizable histologic feature representations to classify the major histopathologic classes of brain tumors. In addition, we implemented a semantic segmentation method to identify tumor-infiltrated diagnostic regions within SRH images. These results demonstrate how intraoperative cancer diagnosis can be streamlined, creating a complementary pathway for tissue diagnosis that is independent of a traditional pathology laboratory.

Published
2020

7. Toward digital histopathological assessment in surgery for central nervous system tumors using stimulated Raman histology

Author

Wadiura, Lisa I., primary, Kiesel, Barbara, additional, Roetzer-Pejrimovsky, Thomas, additional, Mischkulnig, Mario, additional, Vogel, Clemens C., additional, Hainfellner, Johannes A., additional, Matula, Christian, additional, Freudiger, Christian W., additional, Orringer, Daniel A., additional, Wöhrer, Adelheid, additional, Roessler, Karl, additional, and Widhalm, Georg, additional

Published
2022
Full Text
View/download PDF

9. Additional file 1 of Novel rapid intraoperative qualitative tumor detection by a residual convolutional neural network using label-free stimulated Raman scattering microscopy

Author

Reinecke, David, von Spreckelsen, Niklas, Mawrin, Christian, Ion-Margineanu, Adrian, Fürtjes, Gina, Jünger, Stephanie T., Khalid, Florian, Freudiger, Christian W., Timmer, Marco, Ruge, Maximilian I., Goldbrunner, Roland, and Neuschmelting, Volker

Abstract

Additional file 1. Table S1: Internal consistency analysis and overall agreement of the three residual Convolutional Neural Networks (CNNs). Table S2: Comparison of prediction probabilities of the three residual Convolutional Neural Networks (CNN) of the corresponding random areas of the same tumor samples in mean and standard deviation, and Intraclass correlation coefficient (ICC). Table S3: Subdivision of predictions of neuropathological evaluation (NP), residual convolutional neural network (CNN), and each interrater agreement (CP) based on the different entities.

Published
2022
Full Text
View/download PDF

11. Rapid Automated Analysis of Skull Base Tumor Specimens Using Intraoperative Optical Imaging and Artificial Intelligence

Author

Jiang, Cheng, primary, Bhattacharya, Abhishek, additional, Linzey, Joseph R., additional, Joshi, Rushikesh S., additional, Cha, Sung Jik, additional, Srinivasan, Sudharsan, additional, Alber, Daniel, additional, Kondepudi, Akhil, additional, Urias, Esteban, additional, Pandian, Balaji, additional, Al-Holou, Wajd N., additional, Sullivan, Stephen E., additional, Thompson, B. Gregory, additional, Heth, Jason A., additional, Freudiger, Christian W., additional, Khalsa, Siri Sahib S., additional, Pacione, Donato R., additional, Golfinos, John G., additional, Camelo-Piragua, Sandra, additional, Orringer, Daniel A., additional, Lee, Honglak, additional, and Hollon, Todd C., additional

Published
2022
Full Text
View/download PDF

14. Rapid Automated Analysis of Skull Base Tumor Specimens Using Intraoperative Optical Imaging and Artificial Intelligence.

Author

Cheng Jiang, Bhattacharya, Abhishek, Linzey, Joseph R., Joshi, Rushikesh S., Sung Jik Cha, Srinivasan, Sudharsan, Alber, Daniel, Kondepudi, Akhil, Urias, Esteban, Pandian, Balaji, Al-Holou, Wajd N., Sullivan, Stephen E., Thompson, B. Gregory, Heth, Jason A., Freudiger, Christian W., Khalsa, Siri Sahib S., Pacione, Donato R., Golfinos, John G., Camelo-Piragua, Sandra, and Orringer, Daniel A.

Published
2022
Full Text
View/download PDF

15. Rapid, label-free detection of diffuse glioma recurrence using intraoperative stimulated Raman histology and deep neural networks

Author

Hollon, Todd C, primary, Pandian, Balaji, additional, Urias, Esteban, additional, Save, Akshay V, additional, Adapa, Arjun R, additional, Srinivasan, Sudharsan, additional, Jairath, Neil K, additional, Farooq, Zia, additional, Marie, Tamara, additional, Al-Holou, Wajd N, additional, Eddy, Karen, additional, Heth, Jason A, additional, Khalsa, Siri Sahib S, additional, Conway, Kyle, additional, Sagher, Oren, additional, Bruce, Jeffrey N, additional, Canoll, Peter, additional, Freudiger, Christian W, additional, Camelo-Piragua, Sandra, additional, Lee, Honglak, additional, and Orringer, Daniel A, additional

Published
2020
Full Text
View/download PDF

17. Rapid, label-free detection of diffuse glioma recurrence using intraoperative stimulated Raman histology and deep neural networks.

Author

Hollon, Todd C, Pandian, Balaji, Urias, Esteban, Save, Akshay V, Adapa, Arjun R, Srinivasan, Sudharsan, Jairath, Neil K, Farooq, Zia, Marie, Tamara, Al-Holou, Wajd N, Eddy, Karen, Heth, Jason A, Khalsa, Siri Sahib S, Conway, Kyle, Sagher, Oren, Bruce, Jeffrey N, Canoll, Peter, Freudiger, Christian W, Camelo-Piragua, Sandra, and Lee, Honglak

Published
2021
Full Text
View/download PDF

18. Optical Heterodyne-Detected Raman-Induced Kerr Effect (OHD-RIKE) Microscopy

Author

Freudiger, Christian W., Roeffaers, Maarten B. J., Zhang, Xu, Saar, Brian G., Min, Wei, and Xie, X. Sunney

Abstract

Label-free microscopy based on Raman scattering has been increasingly used in biomedical research to image samples that cannot be labeled or stained. Stimulated Raman scattering (SRS) microscopy allows signal amplification of the weak Raman signal for fast imaging speeds without introducing the nonresonant background and coherent image artifacts that are present in coherent anti-Stokes Raman scattering (CARS) microscopy. Here we present the Raman-induced Kerr effect (RIKE) as a contrast for label-free microscopy. RIKE allows us to measure different elements of the nonlinear susceptibility tensor, both the real and imaginary parts, by optical heterodyne detection (OHD-RIKE). OHD-RIKE microscopy provides information similar to polarization CARS (P-CARS) and interferometric CARS (I-CARS) microscopy, with a simple modification of the two-beam SRS microscopy setup. We show that, while OHD-RIKE microspectroscopy can be in principle more sensitive than SRS, it does not supersede SRS microscopy of heterogeneous biological samples, such as mouse skin tissue, because it is complicated by variations of linear birefringence across the sample.

Published
2024
Full Text
View/download PDF

19. Rapid intraoperative histology of unprocessed surgical specimens via fibre-laser-based stimulated Raman scattering microscopy

Author

Orringer, Daniel A., primary, Pandian, Balaji, additional, Niknafs, Yashar S., additional, Hollon, Todd C., additional, Boyle, Julianne, additional, Lewis, Spencer, additional, Garrard, Mia, additional, Hervey-Jumper, Shawn L., additional, Garton, Hugh J. L., additional, Maher, Cormac O., additional, Heth, Jason A., additional, Sagher, Oren, additional, Wilkinson, D. Andrew, additional, Snuderl, Matija, additional, Venneti, Sriram, additional, Ramkissoon, Shakti H., additional, McFadden, Kathryn A., additional, Fisher-Hubbard, Amanda, additional, Lieberman, Andrew P., additional, Johnson, Timothy D., additional, Xie, X. Sunney, additional, Trautman, Jay K., additional, Freudiger, Christian W., additional, and Camelo-Piragua, Sandra, additional

Published
2017
Full Text
View/download PDF

20. SURG-13. STIMULATED RAMAN SCATTERING MICROSCOPY PROVIDES DIAGNOSTIC INTRAOPERATIVE HISTOPATHOLOGIC IMAGES IN BRAIN TUMOR PATIENTS

Author

Orringer, Daniel A., primary, Pandian, Balaji, additional, Hollon, Todd C., additional, Niknafs, Yashar S., additional, Boyle, Julianne, additional, Lewis, Spencer, additional, Hervey-Jumper, Shawn L., additional, Garton, Hugh J.L., additional, Maher, Cormac O., additional, Heth, Jason A., additional, Sagher, Oren, additional, Snuderl, Matija, additional, Venneti, Sriram, additional, Ramkissoon, Shakti, additional, McFadden, Kathryn A., additional, Fisher-Hubbard, Amanda, additional, Lieberman, Andrew, additional, Johnson, Timothy D., additional, Xie, X. Sunney, additional, Freudiger, Christian W., additional, and Camelo-Piragua, Sandra, additional

Published
2016
Full Text
View/download PDF

22. Detection of human brain tumor infiltration with quantitative stimulated Raman scattering microscopy

Author

Ji, Minbiao, primary, Lewis, Spencer, additional, Camelo-Piragua, Sandra, additional, Ramkissoon, Shakti H., additional, Snuderl, Matija, additional, Venneti, Sriram, additional, Fisher-Hubbard, Amanda, additional, Garrard, Mia, additional, Fu, Dan, additional, Wang, Anthony C., additional, Heth, Jason A., additional, Maher, Cormac O., additional, Sanai, Nader, additional, Johnson, Timothy D., additional, Freudiger, Christian W., additional, Sagher, Oren, additional, Xie, Xiaoliang Sunney, additional, and Orringer, Daniel A., additional

Published
2015
Full Text
View/download PDF

24. Rapid, Label-Free Detection of Brain Tumors with Stimulated Raman Scattering Microscopy

Author

Ji, Minbiao, primary, Orringer, Daniel A., additional, Freudiger, Christian W., additional, Ramkissoon, Shakti, additional, Liu, Xiaohui, additional, Lau, Darryl, additional, Golby, Alexandra J., additional, Norton, Isaiah, additional, Hayashi, Marika, additional, Agar, Nathalie Y. R., additional, Young, Geoffrey S., additional, Spino, Cathie, additional, Santagata, Sandro, additional, Camelo-Piragua, Sandra, additional, Ligon, Keith L., additional, Sagher, Oren, additional, and Xie, X. Sunney, additional

Published
2013
Full Text
View/download PDF

26. Erratum: Multicolored stain-free histopathology with coherent Raman imaging

Author

Freudiger, Christian W, primary, Pfannl, Rolf, additional, Orringer, Daniel A, additional, Saar, Brian G, additional, Ji, Minbiao, additional, Zeng, Qing, additional, Ottoboni, Linda, additional, Ying, Wei, additional, Waeber, Christian, additional, Sims, John R, additional, De Jager, Philip L, additional, Sagher, Oren, additional, Philbert, Martin A, additional, Xu, Xiaoyin, additional, Kesari, Santosh, additional, Xie, X Sunney, additional, and Young, Geoffrey S, additional

Published
2012
Full Text
View/download PDF

41. Label-free DNA imaging in vivo with stimulated Raman scattering microscopy.

Author

Fa-Ke Lu, Basu, Srinjan, Igras, Vivien, Hoang, Mai P., Minbiao Ji, Dan Fu, Holtom, Gary R., Neel, Victor A., Freudiger, Christian W., Fisher, David E., and Xie, X. Sunney

Subjects
RAMAN scattering, SKIN cancer -- Genetic aspects, CELL division, CELL cycle, HISTOLOGICAL techniques
Abstract

Label-free DNA imaging is highly desirable in biology and medicine to perform live imaging without affecting cell function and to obtain instant histological tissue examination during surgical procedures. Here we show a label-free DNA imaging method with stimulated Raman scattering (SRS) microscopy for visualization of the cell nuclei in live animals and intact fresh human tissues with subcellular resolution. Relying on the distinct Raman spectral features of the carbon-hydrogen bonds in DNA, the distribution of DNA is retrieved from the strong background of proteins and lipids by linear decomposition of SRS images at three optimally selected Raman shifts. Based on changes on DNA condensation in the nucleus, we were able to capture chromosome dynamics during cell division both in vitro and in vivo.We tracked mouse skin cell proliferation, induced by drug treatment, through in vivo counting of the mitotic rate. Furthermore, we demonstrated a label-free histology method for human skin cancer diagnosis that provides comparable results to other conventional tissue staining methods such as H&E. Our approach exhibits higher sensitivity than SRS imaging of DNA in the fingerprint spectral region. Compared with spontaneous Raman imaging of DNA, our approach is three orders of magnitude faster, allowing both chromatin dynamic studies and label-free optical histology in real time. [ABSTRACT FROM AUTHOR]

Published
2015
Full Text
View/download PDF

45. Coherent Nonlinear Optical Imaging: Beyond Fluorescence Microscopy.

Author

Wei Min, Freudiger, Christian W., Sijia Lu, and Xie, X. Sunney

Subjects
*SPECTRUM analysis, *FLUORESCENCE, *FLUORESCENCE microscopy, *RAMAN spectroscopy, *MICROSCOPY
Abstract

The quest for ultrahigh detection sensitivity with spectroscopic contrasts other than fluorescence has led to various novel approaches to optical microscopy of biological systems. Coherent nonlinear optical imaging, especially the recently developed nonlinear dissipation microscopy (including stimulated Raman scattering and two-photon absorption) and pump-probe microscopy (including excited-state absorption, stimulated emission, and ground-state depletion), provides new image contrasts for nonfluorescent species. Thanks to the high-frequency modulation transfer scheme, these imaging techniques exhibit superb detection sensitivity. By directly interrogating vibrational and/or electronic energy levels of molecules, they offer high molecular specificity. Here we review the underlying principles and excitation and detection schemes, as well as exemplary biomedical applications of this emerging class of molecular imaging techniques. [ABSTRACT FROM AUTHOR]

Published
2011
Full Text
View/download PDF

46. Optical Heterodyne-Detected Raman-Induced Kerr Effect (OHD-RIKE) Microscopy

Author

Freudiger, Christian W., Roeffaers, Maarten B. J., Zhang, Xu, Saar, Brian G., Min, Wei, and Xie, Xiaoliang Sunney

Abstract

Label-free microscopy based on Raman scattering has been increasingly used in biomedical research to image samples that cannot be labeled or stained. Stimulated Raman scattering (SRS) microscopy allows signal amplification of the weak Raman signal for fast imaging speeds without introducing the nonresonant background and coherent image artifacts that are present in coherent anti-Stokes Raman scattering (CARS) microscopy. Here we present the Raman-induced Kerr effect (RIKE) as a contrast for label-free microscopy. RIKE allows us to measure different elements of the nonlinear susceptibility tensor, both the real and imaginary parts, by optical heterodyne detection (OHD-RIKE). OHD-RIKE microscopy provides information similar to polarization CARS (P-CARS) and interferometric CARS (I-CARS) microscopy, with a simple modification of the two-beam SRS microscopy setup. We show that, while OHD-RIKE microspectroscopy can be in principle more sensitive than SRS, it does not supersede SRS microscopy of heterogeneous biological samples, such as mouse skin tissue, because it is complicated by variations of linear birefringence across the sample., Chemistry and Chemical Biology

Published
2011
Full Text
View/download PDF

48. Coherent Raman Tissue Imaging in the Brain

Author

Saar, Brian G., Freudiger, Christian W., Xu, Xiaoyin, Huttner, Anita, Kesari, Santosh, Young, Geoffrey, and Xie, X. Sunney

Abstract

Imaging in neuroscience has been dramatically impacted by the advent of multiphoton microscopy. Multiphoton-excited fluorescence (MPF) in combination with endogenous fluorophores or labeling by fluorescent molecules has proven to be particularly powerful. However, endogenous fluorescence is limited to relatively few molecular species, and practical labeling schemes do not exist for many classes of molecules. Coherent Raman scattering (CRS) techniques, including coherent anti-Stokes Raman scattering and stimulated Raman scattering, allow imaging without the need for staining or fluorescent labeling. Such label-free imaging is desirable in biomedical research, because labeling often perturbs the function of small metabolite and drug molecules and may be too toxic to use in vivo. CRS techniques have similar imaging parameters to MPF, making use of pulsed near-infrared lasers to deliver high-sensitivity, high spatial resolution in three dimensions and rapid image acquisition. In this introduction, we will discuss the basic principles of CRS imaging, present the instrumentation requirements for high-speed CRS imaging, and show an example of imaging brain tumors and healthy tissue based on their intrinsic vibrational signatures. This discussion is intended to introduce the benefits and tradeoffs associated with different CRS techniques and show one example of the powerful capabilities of label-free chemical imaging.

Published
2014
Full Text
View/download PDF

50. Hierarchical discriminative learning improves visual representations of biomedical microscopy.

Author

Jiang C, Hou X, Kondepudi A, Chowdury A, Freudiger CW, Orringer DA, Lee H, and Hollon TC

Abstract

Learning high-quality, self-supervised, visual representations is essential to advance the role of computer vision in biomedical microscopy and clinical medicine. Previous work has focused on self-supervised representation learning (SSL) methods developed for instance discrimination and applied them directly to image patches, or fields-of-view, sampled from gigapixel whole-slide images (WSIs) used for cancer diagnosis. However, this strategy is limited because it (1) assumes patches from the same patient are independent, (2) neglects the patient-slide-patch hierarchy of clinical biomedical microscopy, and (3) requires strong data augmentations that can degrade downstream performance. Importantly, sampled patches from WSIs of a patient's tumor are a diverse set of image examples that capture the same underlying cancer diagnosis. This motivated HiDisc, a data-driven method that leverages the inherent patient-slide-patch hierarchy of clinical biomedical microscopy to define a hierarchical discriminative learning task that implicitly learns features of the underlying diagnosis. HiDisc uses a self-supervised contrastive learning framework in which positive patch pairs are defined based on a common ancestry in the data hierarchy, and a unified patch, slide, and patient discriminative learning objective is used for visual SSL. We benchmark HiDisc visual representations on two vision tasks using two biomedical microscopy datasets, and demonstrate that (1) HiDisc pretraining outperforms current state-of-the-art self-supervised pretraining methods for cancer diagnosis and genetic mutation prediction, and (2) HiDisc learns high-quality visual representations using natural patch diversity without strong data augmentations.

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results