We have successfully processed 4456 infant brain images from multiple sites with various protocols and scanners (Table 1). Here are some feedback from them:
"Just wanted to say, I have tried many different methods for infant brain segmentation in the ~6 month old brain, and none have performed as well as iBeat V.2.0. I was impressed with how iBeat was able to appreciate distinct patches of reversed tissue contrasts. Thank you so much for your work and help. This is truly a break through towards longitudinal brain segmentation over the first year of life."
"What I've seen so far look far more accurate than the infant pipelines I tried previously."
"I am following up as I have just tested subject 1 using infant freesurfer (T1 input), and the surfaces are poor. The tissue segmentation from iBEAT is excellent in comparison to that from freesurfer for this subject."
"Regarding our own in-house tool, it appears because the T2 data was challenging and much higher thickness, whatever testing my colleagues performed was deemed to be poor and they didn’t proceed much further. It seems so far that iBEAT may be the best option for this data."
"I don't have too much to report though other than that they look great! Far more accurate labelling than other tools I've tried"
“The preliminary version of the tools has helped us process 30+ infant subjects at 6 months of age. I am impressed with the outstanding performance of the tools.”"We are very impressed with the results. Thank you so much for making this available!"
"Thank you. Wow these look incredible, very impressive! As good as FreeSurfer on an adult brain.
"Infant freesurfer, but because it only operates on the T1-weighted image which doesn’t have much tissue contrast in younger infants, the results are pretty bad, even for images that I would consider have almost no noise."
"Thanks, I’m looking forward to submitting more infants soon. What a wonderful software you’ve made!"
"We went through these images and were very impressed by the results! Thank you!"
"Thank you so much for processing these data for free. The segmentations look amazing--best I've ever seen. The data all look great especially after you reprocessed. The originally processed files had ACC segmentation errors but the new files all seem to have this resolved. "
"The result of iBeat Cloud segmentation looks great."
“I am very impressed by the segmentation results achieved by your algorithms.”
“We found out iBEAT performs better than other tools in neonatal segmentation.”
“Your computational tools proved very useful in analyzing MRI data in order to build our brain models.”
“We really appreciate your endeavour to provide such kind of free service! We are happy to find the results of tissue segmentation of 6-month old infants quite good!”
"Wow! I am truly impressed! These results are remarkable: we have never seen grey/white matter segmentation this good! Even when we provided you only the T1w images these results are sufficient for our use case!"
"I have finished my analyses of the first two participants with the surfaces you provided (they look great!)"
"The segmentation looks really good!"
"The results of your new segmentation (and brain extraction) look great! I'm very impressed because we've struggled a lot with getting these images processed. iBEAT had been the software that was working the best for us, but the results you sent back are even better than what I was able to get out of the older version of iBEAT. "
"I finally got a chance to show the results of iBEAT cloud to my collaborators. They are all very impressed, and we plan to upload some more subjects to you in the near future. ...Thank you for all your hard work on this new version of iBEAT! It is very impressive."
"Thank you very much for the processing. The segmentation is fantastic. I am quite happy with the segmentation."
"Thank you so much for running our brains through your segmentation pipeline. Your pipeline did a fantastic job segmenting our brains."
"The results are awesome! We previously had some problems with the segmentation results while using other tools, however the iBEAT segmentation doesn't have the same problem and we are very happy with that. Even with some of the brains that we thought would be challenging to process, iBEAT did a good job with the segmentation. We really appreciate all the effort made to develop such an impressive pipeline!"
"I have to say that my impression about the performance of the software is, by far, better than expected. Actually, I thought it was not going to be possible for the last patients. I am quite happy with the results. Just for make it clear, results are perfect for me and are even better that the dHCP-derived results which is optimized for neonates."
"The segmentations look really good! Huge improvement from the first iBEAT. Thanks so much for processing these data for free!"
"Many thanks for helping on processing data! The results look great! The segmentation results produced by your algorithms are very impressive, and outperformed the other methods we have tried. We will have more data coming, and wish to your further assist on processing. "
"The preliminary version of the tools has helped me process 48 infant subjects. I am very impressed at how well the new version of iBEAT has segmented our newborn data, even in infants with only one image type collected."
"Everything looks great! Thank you very much! "
Table 1. Successfully processed 4456 infant brain images from multiple sites with various protocols and scanners.
iBEAT V2.0 is a toolbox for processing infant brain MR images, using multimodality (including T1w and T2w) or single-modality. The software is developed by the IDEA group at the University of North Carolina at Chapel Hill. iBEAT was first developed in 2012, now re-developed with more advanced techniques.