no code implementations • 20 Sep 2023 • Sivaramakrishnan Rajaraman, Ghada Zamzmi, Feng Yang, Zhaohui Liang, Zhiyun Xue, Sameer Antani
Model initialization techniques are vital for improving the performance and reliability of deep learning models in medical computer vision applications.
no code implementations • 18 Sep 2023 • Sivaramakrishnan Rajaraman, Ghada Zamzmi, Feng Yang, Zhaohui Liang, Zhiyun Xue, Sameer Antani
Deep learning (DL) has demonstrated its innate capacity to independently learn hierarchical features from complex and multi-dimensional data.
no code implementations • 10 Jan 2023 • Sivaramakrishnan Rajaraman, Feng Yang, Ghada Zamzmi, Zhiyun Xue, Sameer Antani
Literature is sparse in discussing the optimal image resolution to train these models for segmenting the Tuberculosis (TB)-consistent lesions in CXRs.
no code implementations • 4 Nov 2022 • Sivaramakrishnan Rajaraman, Feng Yang, Ghada Zamzmi, Zhiyun Xue, Sameer Antani
In this work, our goal is to (i) analyze the generalizability of deep adult lung segmentation models to the pediatric population and (ii) improve performance through a stage-wise, systematic approach consisting of CXR modality-specific weight initializations, stacked ensembles, and an ensemble of stacked ensembles.
no code implementations • 13 Jun 2022 • Sivaramakrishnan Rajaraman, Feng Yang, Ghada Zamzmi, Peng Guo, Zhiyun Xue, Sameer K Antani
We observed that the stacking ensemble demonstrated superior segmentation performance (Dice score: 0. 5743, 95% confidence interval: (0. 4055, 0. 7431)) compared to the individual constituent models and other ensemble methods.
no code implementations • 5 Nov 2021 • Sivaramakrishnan Rajaraman, Gregg Cohen, Lillian Spear, Les Folio, Sameer Antani
It is observed that the bone suppression model ensemble outperformed the individual models in terms of MS-SSIM and other metrics.
no code implementations • 29 Sep 2021 • Sivaramakrishnan Rajaraman, Ghada Zamzmi, Sameer Antani
Currently, the cross-entropy loss remains the de-facto loss function for training deep learning classifiers.
no code implementations • 29 Sep 2021 • Sivaramakrishnan Rajaraman, Prasanth Ganesan, Sameer Antani
In this study, we perform a systematic analysis of the effect of model calibration on its performance on two medical image modalities, namely, chest X-rays and fundus images, using various deep learning classifier backbones.
no code implementations • 9 Apr 2021 • Sivaramakrishnan Rajaraman, Ghada Zamzmi, Les Folio, Philip Alderson, Sameer Antani
However, the presence of bony structures such as ribs and clavicles can obscure subtle abnormalities, resulting in diagnostic errors.
no code implementations • 21 Feb 2021 • Sivaramakrishnan Rajaraman, Les Folio, Jane Dimperio, Philip Alderson, Sameer Antani
We observe that our augmented training strategy helped the CXR modality-specific U-Net models achieve superior performance with test data derived from the TBX11K CXR training distribution as well as from cross-institutional collections (p < 0. 05).
no code implementations • 19 Jun 2020 • Ghada Zamzmi, Sivaramakrishnan Rajaraman, Sameer Antani
We explore different fine-tuning strategies to demonstrate the impact of the strategy on the performance of target medical image tasks.
no code implementations • 16 Apr 2020 • Sivaramakrishnan Rajaraman, Jen Siegelman, Philip O. Alderson, Lucas S. Folio, Les R. Folio, Sameer K. Antani
We demonstrate use of iteratively pruned deep learning model ensembles for detecting pulmonary manifestation of COVID-19 with chest X-rays.