Unsupervised Object Segmentation by Redrawing

By Mickaël Chen, Thierry Artières, Ludovic Denoyer. 2019

148 citations recorded by the source corpus.

Object segmentation is a crucial problem that is usually solved by using supervised learning approaches over very large datasets composed of both images and corresponding object masks. Since the masks have to be provided at pixel level, building such a dataset for any new domain can be very time-consuming. We present ReDO, a new model able to extract objects from images without any annotation in an unsupervised way. It relies on the idea that it should be possible to change the textures or colors of the objects without changing the overall distribution of the dataset. Following this assumption, our approach is based on an adversarial architecture where the generator is guided by an input sample: given an image, it extracts the object mask, then redraws a new object at the same location. The generator is controlled by a discriminator that ensures that the distribution of generated images is aligned to the original one. We experiment with this method on different datasets and demonstrate the good quality of extracted masks.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Object-Centric Learning with Slot Attention (2020, background): “[48] Mickaël Chen, Thierry Artières, and Ludovic Denoyer.”
GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations (2019, background): “While Chen et al. [22] and Bielski and Favaro [23] segment a single foreground object per image, Arandjelovi´c and Zisserman [24] are able to segment several synthetic objects superimposed on natural images.”
A Metaverse: Taxonomy, Components, Applications, and Open Challenges (2022, unknown): “In particular, some studies (e.g., world models and MONET) abstract multiple objects into representations for fast object recognition and efﬁcient training [77].”
BlockGAN: Learning 3D Object-aware Scene Representations from Unlabelled Images (2020, background): “Moreover, they also assume that each object’s appearance is independent [3, 6, 12].”
Labels4Free: Unsupervised Segmentation using StyleGAN (2021, background): “In the GAN domain, some works [7, 11] formulate multiple tasks in unsupervised settings.”
Development and Application of Self-Supervised Machine Learning for Smoke Plume and Active Fire Identification from the FIREX-AQ Datasets (2025, unknown): “Work on the general problem of self-supervised image segmentation appears to have had success in separating the foreground from the background [22,23], or have only used single bands of input from one type of instrumentation, which is effective for their applications, but does not cover the breadth…”