In-Context Adaptation of VLMs for Few-Shot Cell Detection in Optical Microscopy

Shreyan Ganguly, Angona Biswas, Jaydeep Rade, Md Hasibul Hasan Hasib, Nabila Masud,Nitish Singla, Abhipsa Dash, Ushashi Bhattacharjee, Aditya Balu, Anwesha Sarkar, Adarsh Krishnamurthy, Soumik Sarkar · 2025

Read In-Context Adaptation of VLMs for Few-Shot Cell Detection in Optical Microscopy by Shreyan Ganguly, Angona Biswas, Jaydeep Rade, Md Hasibul Hasan Hasib, Nabila Masud,Nitish Singla, Abhipsa Dash, Ushashi Bhattacharjee, Aditya Balu, Anwesha Sarkar, Adarsh Krishnamurthy, Soumik Sarkar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-05565.txt.

Description

Shreyan Ganguly and a team of co-authors test whether vision-language models can find cells in microscope images from just a few examples. Labs often lack the large marked data sets that detection models need. The team built Micro-OD, a benchmark of 252 images with boxes drawn around eleven cell types from four sources.

They tested eight models with few-shot prompts, meaning the model sees a handful of examples first. Zero-shot results were weak because microscope images differ so much from normal photos. A few examples helped a lot, with gains leveling off after about six. A hybrid pipeline that pairs a detector with a model-based classifier did best.

Catalog record

CREATORShreyan Ganguly, Angona Biswas, Jaydeep Rade, Md Hasibul Hasan Hasib, Nabila Masud,Nitish Singla, Abhipsa Dash, Ushashi Bhattacharjee, Aditya Balu, Anwesha Sarkar, Adarsh Krishnamurthy, Soumik Sarkar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-05565

Browse the full library · About the Tower of Babel