An Open-Source Reproduction and Enhancement of CheXNet for Chest X-ray Disease Classification

[dstrick17]

Original article:
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
Rajpurkar et al., 2017

PDF URL:
An Open-Source Reproduction and Enhancement of CheXNet for Chest X-ray Disease Classification

Code URL:
https://github.com/dstrick17/DacNet

Scientific domain:
Computational Medicine; Medical Imaging; Machine Learning

Programming language:
Python

Suggested editor:
No preference.

[dstrick17]

Hi ReScience Editors, I am checking in to see if any further information is required to move this submission forward to the editor assignment stage. We have the code and reproduction environment ready for review. Thank you for your time.

[rougier]

Oups, really sorry. Your request slipped through. Thanks you for your submission. However it is not clear if this is a replication since you did not provide links to your code nor article. Can you provide them before we move forward in the process?

[dstrick17]

Hi @rougier, thank you for the response! To clarify the submission, here are the links to our reproduction work:

Reproduction Code: [https://github.com/dstrick17/DacNet]

Reproduction Article (PDF): [https://arxiv.org/abs/2505.06646]

This study aims to be a reproduction of the 121-layer DenseNet architecture described in CheXNet using the NIH ChestX-ray14 dataset, with an additional enhancement section introducing our DACNet model. We have ensured a patient-wise split to maintain the integrity of the reproduction.

Please let us know if you need any other metadata to proceed!

[rougier]

Thanks for the precision I've edited your initial submission.
@gdetor Can handle this submission?

[gdetor]

HI @rougier Sure, I can handle this.