brightfield vs ihc, one does not contradict the other. they coexist.
this is a misconception that i had that let me to believe that my enemy is advanced destructive assays.
destructive assays like IHC are still the gold standard. they are probably not going away until digital twins are created (actually that might not even be true). but labs do it for endpoint analysis, but when it's going from EBs to spheroids to BOs, brightfield analysis is still happening. happening in an eyeballing way.
so one doesn't contradict the other. they complement each other. and my enemy isn't destructive assays. my enemy is other non-invasive but expensive methods.
i have to bet against hardware advancements with data. the more data i get the more software alone would become more robust. the goal is to achieve high accuracy with software alone.
updates: umass filled the form! so did one of the study groups in gage lab of salk. 1300+ images acquired. meeting nidi tmr to discuss project commitment and project advancements regarding outreach, comp bio cto search, and objective alignment. updated nick about my progress as well.
i need to get more data and rely on ZERO augmentation.
everything hinges on data. data is the oil of ML.