AI X-Ray Bone Tumor Detection
Budget: ₹750 – ₹1,250 INR
I need a deep-learning model that can take a standard digital X-ray and reliably flag the presence of bone tumors. The core of the project is pure image analysis: preprocessing the radiographs, localising suspicious regions, classifying benign vs malignant patterns, and outputting a confidence score that radiologists can review.
Please work with publicly available, properly anonymised datasets first; I will supply additional proprietary X-rays once the pipeline is in place. Python with TensorFlow or PyTorch is preferred so that my in-house engineers can maintain the code later, and the final network should run on a single mid-range GPU without exotic dependencies.
Deliverables
• End-to-end training notebook or scripts
• Inference API or lightweight GUI that loads an X-ray and returns heatmaps plus a binary tumor/no-tumor verdict
• Model weights and clear instructions for retraining or fine-tuning on new data
• Short report detailing data splits, augmentation strategy, and performance metrics (AUC, sensitivity, specificity) validated on a held-out test set
Acceptance criteria: minimum AUC 0.90 on the test set and inference time under two seconds per image on an RTX 3060.
Please work with publicly available, properly anonymised datasets first; I will supply additional proprietary X-rays once the pipeline is in place. Python with TensorFlow or PyTorch is preferred so that my in-house engineers can maintain the code later, and the final network should run on a single mid-range GPU without exotic dependencies.
Deliverables
• End-to-end training notebook or scripts
• Inference API or lightweight GUI that loads an X-ray and returns heatmaps plus a binary tumor/no-tumor verdict
• Model weights and clear instructions for retraining or fine-tuning on new data
• Short report detailing data splits, augmentation strategy, and performance metrics (AUC, sensitivity, specificity) validated on a held-out test set
Acceptance criteria: minimum AUC 0.90 on the test set and inference time under two seconds per image on an RTX 3060.