Program-Guided Image Manipulators
This addresses image manipulation tasks for computer vision applications, offering a unified framework that learns from single images, but it appears incremental as it builds on existing neuro-symbolic and inpainting methods.
The paper tackles the problem of representing and manipulating images by inducing neuro-symbolic program-like representations to detect patterns and edit images, achieving superior performance on tasks like inpainting, extrapolation, and regularity editing.
Humans are capable of building holistic representations for images at various levels, from local objects, to pairwise relations, to global structures. The interpretation of structures involves reasoning over repetition and symmetry of the objects in the image. In this paper, we present the Program-Guided Image Manipulator (PG-IM), inducing neuro-symbolic program-like representations to represent and manipulate images. Given an image, PG-IM detects repeated patterns, induces symbolic programs, and manipulates the image using a neural network that is guided by the program. PG-IM learns from a single image, exploiting its internal statistics. Despite trained only on image inpainting, PG-IM is directly capable of extrapolation and regularity editing in a unified framework. Extensive experiments show that PG-IM achieves superior performance on all the tasks.