arXiv subjects

Satyaki Roy Chowdhury

Publications and source records attributed to Satyaki Roy Chowdhury.

3 recordsLinked to original sources

arXiv2026

From Bands to Depth: Understanding Bathymetry Decisions on Sentinel-2

Deploying Sentinel-2 satellite derived bathymetry (SDB) robustly across sites remains challenging. We analyze a Swin-Transformer based U-Net model (Swin-BathyUNet) to understand how it infers depth and when its predictions are trustworthy. A leave-one-band out study ranks spectral importance to the different bands consistent with shallow water optics. We adapt ablation-based CAM to regression (A-CAM-R) and validate the reliability via a performance retention test: keeping only the top-p% salient pixels while neutralizing the rest causes large, monotonic RMSE increase, indicating explanations localize on evidence the model relies on. Attention ablations show decoder conditioned cross attention on skips is an effective upgrade, improving robustness to glint/foam. Cross-region inference (train on one site, test on another) reveals depth-dependent degradation: MAE rises nearly linearly with depth, and bimodal depth distributions exacerbate mid/deep errors. Practical guidance follows: maintain wide receptive fields, preserve radiometric fidelity in green/blue channels, pre-filter bright high variance near shore, and pair light target site fine tuning with depth aware calibration to transfer across regions.

cs.CV↗

arXiv2026

Variance-Penalized MC-Dropout as a Learned Smoothing Prior for Brain Tumour Segmentation

Brain tumor segmentation is essential for diagnosis and treatment planning, yet many CNN and U-Net based approaches produce noisy boundaries in regions of tumor infiltration. We introduce UAMSA-UNet, an Uncertainty-Aware Multi-Scale Attention-based Bayesian U-Net that in- stead leverages Monte Carlo Dropout to learn a data-driven smoothing prior over its predictions, while fusing multi-scale features and attention maps to capture both fine details and global context. Our smoothing-regularized loss augments binary cross-entropy with a variance penalty across stochas- tic forward passes, discouraging spurious fluctuations and yielding spatially coherent masks. On BraTS2023, UAMSA- UNet improves Dice Similarity Coefficient by up to 3.3% and mean IoU by up to 2.7% over U-Net; on BraTS2024, it delivers up to 4.5% Dice and 4.0% IoU gains over the best baseline. Remarkably, it also reduces FLOPs by 42.5% rel- ative to U-Net++ while maintaining higher accuracy. These results demonstrate that, by combining multi-scale attention with a learned smoothing prior, UAMSA-UNet achieves both better segmentation quality and computational efficiency, and provides a flexible foundation for future integration with transformer-based modules for further enhanced segmenta- tion results.

cs.CV↗

arXiv2025

Hybridization of Attention UNet with Repeated Atrous Spatial Pyramid Pooling for Improved Brain Tumour Segmentation

Brain tumors are highly heterogeneous in terms of their spatial and scaling characteristics, making tumor segmentation in medical images a difficult task that might result in wrong diagnosis and therapy. Automation of a task like tumor segmentation is expected to enhance objectivity, repeatability and at the same time reducing turn around time. Conventional convolutional neural networks (CNNs) exhibit sub-par performance as a result of their inability to accurately represent the range of tumor sizes and forms. Developing on that, UNets have been a commonly used solution for semantic segmentation, and it uses a downsampling-upsampling approach to segment tumors. This paper proposes a novel architecture that integrates Attention-UNet with repeated Atrous Spatial Pyramid Pooling (ASPP). ASPP effectively captures multi-scale contextual information through parallel atrous convolutions with varying dilation rates. This allows for efficient expansion of the receptive field while maintaining fine details. The attention provides the necessary context by incorporating local characteristics with their corresponding global dependencies. This integration significantly enhances semantic segmentation performance. Our approach demonstrates significant improvements over UNet, Attention UNet and Attention UNet with Spatial Pyramid Pooling allowing to set a new benchmark for tumor segmentation tasks.

eess.IV↗