Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.
Skip to main content

Table 1 mAP50 scores in validation and test sets for models with varying pre-training methods at different input resolutions

From: BarlowTwins-CXR: enhancing chest X-ray abnormality localization in heterogeneous data with cross-domain self-supervised learning

BackBone_weight

Input_size

mAP50 (val set)

mAP50 (test set)

Baseline No Pre-training

224

0.1388 (0.1337, 0.1439)

0.1342 (0.1291, 0.1393)

Pre-trained on ImageNet

 

0.2245 (0.2187, 0.2303)

0.2210 (0.2187, 0.2233)

Random initialization + Barlow twins

 

0.2555 (0.2456, 0.2655)

0.2448 (0.2400, 0.2496)

Barlow Twins + ImageNet pretrained

 

0.2625 (0.2544, 0.2706)

0.2502 (0.2465, 0.2539)

ImageNet pretrained

640

0.2973 (0.2876, 0.3061)

0.2800 (0.2729, 0.2876)

Barlow Twins + ImageNet pretrained

 

0.3102 (0.3076, 0.3139)

0.2890 (0.2787, 0.2993)

  1. Scores are presented with 95% confidence intervals
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us