APAdarsh Bhandary Panambur

Adarsh Bhandary Panambur

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Adarsh Bhandary Panambur has a total of 26 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    BE-WISER: Class-Aware Weak Slice-Level Supervision for Breast MRI

    by Adarsh Bhandary Panambur, Tri-Thien Nguyen, Saahil Islam, Michael Uder, Sebastian Bickelhaupt, Andreas Maier, Siming Bayer

    Purpose: Breast MRI is sensitive for cancer assessment, but deep learning models often require dense lesion annotations. This study evaluates whether weak slice-level supervision can be extended from binary lesion localization to class-aware slice semantics. Methods: We propose BE-WISER, a weakly supervised framework that builds on the recently proposed BE-WISE approach by extending binary slice-level lesion localization to three-class slice semantics. BE-WISER models each slice as normal, benign, or malignant and combines a Swin Transformer slice encoder, attention-based multiple-instance learning for breast-level classification, and a class-aware slice head for weak lesion localization. Gaussian targets from a single radiologist-indicated lesion slice provide weak supervision, and radiologist review was used to assess the clinical plausibility of predicted high-probability slices. Results: On the independent ODELIA test set, BE-WISER with focal loss achieved the best breast-level performance, with an AUC of 0.864 and an ODELIA score of 0.699. This improved over BE-WISE with focal loss by 1.2 and 2.6 percentage points, respectively. For localization, BE-WISER achieved the lowest mean slice distance of 2.505 slices and the highest Hit@1 of 0.632. Radiologist review indicated that highprobability slices corresponded well to lesion-relevant slices. Conclusion: BE-WISER demonstrates that pathology-aware slice semantics improve weakly supervised breast MRI diagnosis and lesion localization while preserving clinically interpretable slice-level evidence.

  • Opportunistic Breast Cancer Risk Stratification From Low-dose Chest CT Using Multiple Instance Learning

    by Yaqiong Ni, Adarsh Bhandary Panambur, Chang Liu, Tri-Thien Nguyen, Siming Bayer, Huang Juan, Sun Jiayu, Lv Su, Andreas Maier

  • Breast MRI Evaluation with Weakly-informed Slice-level Explanation

    by Adarsh Bhandary Panambur, Tri-Thien Nguyen, Siming Bayer, Andreas Maier

  • Open Access

    Patch-CLIP - Contrastive Health Record-Image Joint Training with Patch Embedding Loss

    by Sheethal Bhat, Awais Mansoor, Bogdan Georgescu, Mathias Zinnen, Pranjal Sahu, Adarsh B. Panambur, Florin C. Ghesu, Sasa Grbic, Andreas Maier

    Vision-Language (VL) models such as Contrastive Language-Image pretraining (CLIP) use multimodal self-supervised learning (SSL) methods to extract maximal information from large-scale datasets. This enables the trained model to learn key image encodings while correlating them to corresponding textual information through a contrastive loss function that maximizes the similarity of VL pairs. Due to weak supervision provided by the text information, these VL models demonstrate strong zero-shot classification performance; however, their performance on downstream tasks such as object detection and localization remains suboptimal [1]. In this work, we introduce a novel contrastive loss function that aligns image patch embeddings with text embeddings. These patch embeddings that are usually discarded output from the image encoder can naturally incorporate location information during unsupervised training. The proposed approach improves both localization and classification performance, thus allowing key findings to be localized without the need for a complex downstream object detection framework. We evaluated the performance of our proposed method on two chest X-ray (CXR) datasets for abnormality detection and localization tasks. The experiments achieved state-of-the-art (SOTA) results for 8 abnormality detection tasks. Moreover, the patch prediction maps introduced in work considerably reduce False Positive (FP) rates at a given sensitivity, compared to saliency maps.


  • Open Access

    Attention-guided erasing for enhanced transfer learning in breast abnormality classification

    by Adarsh Bhandary Panambur, Sheethal Bhat, Hui Yu, Prathmesh Madhu, Siming Bayer, Andreas Maier

    Purpose: Breast cancer remains one of the most prevalent cancers globally, necessitating effective early screening and diagnosis. This study investigates the effectiveness and generalizability of our recently proposed data augmentation technique, attention-guided erasing (AGE), across various transfer learning classification tasks for breast abnormality classification in mammography. Methods: AGE utilizes attention head visualizations from DINO self-supervised pretraining to weakly localize regions of interest (ROI) in images. These localizations are then used to stochastically erase non-essential background information from training images during transfer learning. Our research evaluates AGE across two image-level and three patch-level classification tasks. The image-level tasks involve breast density categorization in digital mammography (DM) and malignancy classification in contrast-enhanced mammography (CEM). Patch-level tasks include classifying calcifications and masses in scanned film mammography (SFM), as well as malignancy classification of ROIs in CEM. Results: AGE significantly boosts classification performance with statistically significant improvements in mean F1-scores across four tasks compared to baselines. Specifically, for image-level classification of breast density in DM and malignancy in CEM, we achieve gains of 2% and 1.5%, respectively. Additionally, for patch-level classification of calcifications in SFM and CEM ROIs, gains of 0.4% and 0.6% are observed, respectively. However, marginal improvement is noted in the mass classification task, indicating the necessity for further optimization in tasks where critical features may be obscured by erasing techniques. Conclusion: Our findings underscore the potential of AGE, a dataset- and task-specific augmentation strategy powered by self-supervised learning, to enhance the downstream classification performance of DL models, particularly involving ViTs, in medical imaging.


  • Open Access

    Attention-Guided Erasing for Enhanced Transfer Learning in Breast Abnormality Classification

    by Adarsh Bhandary Panambur, Sheethal Bhat, Hui Yu, Prathmesh Madhu, Siming Bayer, Andreas Maier

    Purpose: Breast cancer remains one of the most prevalent cancers globally, necessitating effective early screening and diagnosis. This study investigates the effectiveness and generalizability of our recently proposed data augmentation technique, attention-guided erasing (AGE), across various transfer learning classification tasks for breast abnormality classification in mammography. Methods: AGE utilizes attention head visualizations from DINO self-supervised pretraining to weakly localize regions of interest (ROI) in images. These localizations are then used to stochastically erase non-essential background information from training images during transfer learning. Our research evaluates AGE across two image-level and three patch-level classification tasks. The image-level tasks involve breast density categorization in digital mammography (DM) and malignancy classification in contrast-enhanced mammography (CEM). Patch-level tasks include classifying calcifications and masses in scanned film mammography (SFM), as well as malignancy classification of ROIs in CEM. Results: AGE significantly boosts classification performance with statistically significant improvements in mean F1-scores across four tasks compared to baselines. Specifically, for image-level classification of breast density in DM and malignancy in CEM, we achieve gains of 2% and 1.5%, respectively. Additionally, for patch-level classification of calcifications in SFM and CEM ROIs, gains of 0.4% and 0.6% are observed, respectively. However, marginal improvement is noted in the mass classification task, indicating the necessity for further optimization in tasks where critical features may be obscured by erasing techniques. Conclusion: Our findings underscore the potential of AGE, a dataset- and task-specific augmentation strategy powered by self-supervised learning, to enhance the downstream classification performance of deep learning models, particularly involving Vision Transformers, in medical imaging.

  • Attention-guided Erasing

    by Adarsh Bhandary Panambur, Hui Yu, Sheethal Bhat, Prathmesh Madhu, Siming Bayer, Andreas Maier


  • Open Access

    Attention-Guided Erasing: A Novel Augmentation Method for Enhancing Downstream Breast Density Classification

    by Adarsh Bhandary Panambur, Hui Yu, Sheethal Bhat, Prathmesh Madhu, Siming Bayer, Andreas Maier

    The assessment of breast density is crucial in the context of breast cancer screening, especially in populations with a higher percentage of dense breast tissues. This study introduces a novel data augmentation technique termed Attention-Guided Erasing (AGE), devised to enhance the downstream classification of four distinct breast density categories in mammography following the BI-RADS recommendation in the Vietnamese cohort. The proposed method integrates supplementary information during transfer learning, utilizing visual attention maps derived from a vision transformer backbone trained using the self-supervised DINO method. These maps are utilized to erase background regions in the mammogram images, unveiling only the potential areas of dense breast tissues to the network. Through the incorporation of AGE during transfer learning with varying random probabilities, we consistently surpass classification performance compared to scenarios without AGE and the traditional random erasing transformation. We validate our methodology using the publicly available VinDr-Mammo dataset. Specifically, we attain a mean F1-score of 0.5910, outperforming values of 0.5594 and 0.5691 corresponding to scenarios without AGE and with random erasing (RE), respectively. This superiority is further substantiated by t-tests, revealing a p-value of p<0.0001, underscoring the statistical significance of our approach.

  • Open Access

    Effect of Random Histogram Equalization on Breast Calcification Analysis Using Deep Learning

    by Adarsh Bhandary Panambur, Prathmesh Madhu, Andreas Maier

    Early detection and analysis of calcifications in mammogram images is crucial in a breast cancer diagnosis workflow. Management of calcifications that require immediate follow-up and further analyzing its benignancy or malignancy can result in a better prognosis. Recent studies have shown that deep learning-based algorithms can learn robust representations to analyze suspicious calcifications in mammography. In this work, we demonstrate that randomly equalizing the histograms of calcification patches as a data augmentation technique can significantly improve the classification performance for analyzing suspicious calcifications. We validate our approach by using the CBIS-DDSM dataset for two classification tasks. The results on both the tasks show that the proposed methodology gains more than 1% mean accuracy and F1-score when equalizing the data with a probability of 0.4 when compared to not using histogram equalization. This is further supported by the t-tests, where we obtain a p-value of p<0.0001, thus showing the statistical significance of our approach.