🎉 Excited to share that our paper has been published at IEEE TENSYMP held from Jul 7-9 2025 , at Christchurch , New Zealand 🎉 "HybridNet: A CNN-Transformer Based Approach for Myocardial Infarction Segmentation" Authors: Satya Karthik Achanta , Rithwik Eluri, Rama Rakshith Katta, Avasarala Nirupam Laxmi Srujan, Aneesh Narayan Bandaru, and Dr. Santosh Kumar State-of-the-Art Results: Dice Coefficient: 0.985 Weighted IoU: 0.971 F1 Score: 0.964 Classification Accuracy: 0.983 Innovation: Our HybridNet combines the power of U-Net's multi-scale feature extraction with Transformer's global attention mechanisms for accurate myocardial infarction segmentation from cardiac MRI images. Clinical Impact: 130ms inference time per image - suitable for real-time clinical deployment Addresses critical challenges: limited labeled data, complex cardiac motion, computational efficiency Outperforms traditional models (U-Net, SegNet, ResNet) and recent literature Technical Contributions: Novel hybrid CNN-Transformer architecture Advanced preprocessing pipeline with noise reduction and adaptive histogram equalization Combined Dice Loss + Cross-Entropy Loss for optimal segmentation Robust training pipeline with dynamic learning rate scheduling This work represents a significant step forward in automated cardiac diagnosis, potentially enabling faster MI detection and improved patient outcomes, especially in resource-constrained clinical settings. 📖 Read the full paper: https://lnkd.in/dKgksvBn #MachineLearning #MedicalAI #DeepLearning #CardiacImaging #ComputerVision #Healthcare #AI #Research #IEEE #Transformers #CNN #MyocardialInfarction #MedicalImageAnalysis
Incredible work!💐
congratulations 👏
incredible work srujan!!
Amazon ML Summer School '25 | Former Summer Intern @ S&P Global | Final Year CSE-AIML Student @ VNRVJIET, Hyderabad | Web Dev, Data Analysis, ML, AI Enthusiast
5dCongratulations Srujan!!