Instance Separation Emerges from Inpainting
This work addresses instance segmentation for microscopy imaging, offering a self-supervised approach that is incremental in its application of inpainting techniques.
The paper tackled the problem of instance separation in microscopy images by leveraging self-supervised inpainting networks to measure region independence, achieving segmentation performance comparable to fully supervised methods.
Deep neural networks trained to inpaint partially occluded images show a deep understanding of image composition and have even been shown to remove objects from images convincingly. In this work, we investigate how this implicit knowledge of image composition can be leveraged for fully self-supervised instance separation. We propose a measure for the independence of two image regions given a fully self-supervised inpainting network and separate objects by maximizing this independence. We evaluate our method on two microscopy image datasets and show that it reaches similar segmentation performance to fully supervised methods.