Papers/2608.21427
🧪 Test?View on arXiv

Few-Shot Cross-Dataset Adaptation for Tuberculosis Detection Using DenseNet

Author1, Author2, Author3, Author4, Author5

fine-tuningdomain adaptationfew-shot learning
2608.21427
Builder Relevance
80%
3h ago

Abstract

This paper addresses the challenge of domain adaptation in tuberculosis detection using few-shot learning techniques.

Reality Card

Core Claim

The study demonstrates that full fine-tuning of a pretrained DenseNet121 model can achieve 98.36% accuracy with just 75 labeled samples per class, effectively mitigating domain shift.

Method / Result

Achieved 98.36% accuracy with only 75 labeled samples per class.

Limitations

The study may face challenges in reproducibility due to variations in imaging protocols and patient demographics across datasets.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers