Top suggestions for Vision Language Model in Medical ImagingExplore more searches like Vision Language Model in Medical Imaging |
+have+achieved+significant+success+in+recent+years%252C+and+they+have+been+extended+to+the+medical+domain.+Although+demonstrating+satisfactory+performance+on+medical+Visual+Question+Answering+(VQA)+tasks%252C+Medical+LVLMs+(MLVLMs)+suffer+from+the+hallucination+problem%252C+which+makes+them+fail+to+diagnose+complex+pathologies.+Moreover%252C+they+readily+fail+to+learn+minority+pathologies+due+to+imbalanced+training+data.+We+propose+two+prompting+strategies+for+MLVLMs+that+reduce+hallucination+and+improve+VQA+performance.+In+the+first+strategy%252C+we+provide+a+detailed+explanation+of+the+queried+pathology.+In+the+second+strategy%252C+we+fine-tune+a+cheap%252C+weak+learner+to+achieve+high+performance+on+a+specific+metric%252C+and+textually+provide+its+judgment+to+the+MLVLM.+Tested+on+the+MIMIC-CXR-JPG+and+Chexpert+datasets%252C+our+methods+significantly+improve+the+diagnostic+F1+score%252C+with+the+highest+increase+being+0.27.+We+also+demonstrate+that+our+prompting+strategies+can+be+extended+to+general+LVLM+domains.+Based+on+POPE+metrics%252C+it+effectively+suppresses+the+false+negative+predictions+of+existing+LVLMs+and+improves+Recall+by+approximately+0.07.&cdnurl=https%3a%2f%2fth.bing.com%2fth%2fid%2fR.5153b345f897a30bae7cd5b1548bb4a2%3frik%3dDpQVNKDK0ckzPQ%26pid%3dImgRaw%26r%3d0&exph=630&expw=1200&q=Vision+Language+Model+in+Medical+Imaging&FORM=IRPRST&ck=0943CF9A20FCBB1845BF006D60260E21&selectedIndex=2&itb=0)
Explore more searches like Vision Language Model
in Medical Imaging
Some results have been hidden because they may be inaccessible to you.Show inaccessible results
Feedback
