Bioimage Classification with Handcrafted and Learned Features

摘要: Bioimage classification is increasingly becoming more important in many biological studies including those that require accurate cell phenotype recognition, subcellular localization, and histopathological classification. In this paper, we present a new General Purpose (GenP) bioimage method can be applied to large range of problems. The GenP system propose an ensemble combines multiple texture features (both handcrafted learned descriptors) for superior generalizable discriminative power. Our obtains boosting performance by combining local features, dense sampling deep learning features. Each descriptor used train different Support Vector Machine then combined sum rule. We evaluate our on diverse set tasks each represented benchmark database, some available the IICBU 2008 database. task represents typical subcellular, cellular, tissue level problem. evaluation these datasets demonstrates proposed state-of-the-art without any ad-hoc dataset tuning parameters (thereby avoiding risk overfitting/overtraining). To reproduce experiments reported MATLAB code all descriptors at https://github.com/LorisNanni https://www.dropbox.com/s/bguw035yrqz0pwp/ElencoCode.docx?dl=0.