Abstract
Instance-level image translation aims to only translate instance of interest and can be operated more finely and flexibly than object-level and holistic-level image translation. However, current algorithms are not suitable to do it since they employ a holistic or object level’s discriminator that tends to change the whole image or all instances. To address the issue, we propose a simple yet effective local discriminator, in which the input image is split into two parts, region of interest (ROI) and background. Instance mask is employed to align the ROI and the background is design to be random in a prior distribution to mitigate a divergence between the ROI and the background. In this way, we obtain translated instance with decent margins without artifacts as current algorithms get. Moreover we propose a new architecture to simultaneously realize versatile instance-level image translation. Experimental results prove that our proposed algorithm outperforms the state-of-the-art in position accuracy and background retainment by a clear margin.
Highlights
A rapid development has been witnessed in image-toimage (I2I) translation with generative adversarial networks (GAN) [1]
PRELIMINARY As defined before, the instance-level image translation in this paper aims to translate the instance to a specific domain but retain the background, which differs from the usage that instance-level is considered but for holistic image translation, [15], [13], [14]
To achieve instance-level image translation which requires to translate the given instance and retain the background, we proposed a local discriminator and a versatile generator in this paper
Summary
A rapid development has been witnessed in image-toimage (I2I) translation with generative adversarial networks (GAN) [1]. In holistic-level image translation, holistic image is taken as one domain and expected to be translated, such as style transferring [7] and semantic image synthesis [8], [9] but one natural images commonly include various objects. Object-level image translation [10], [11], [12], [4], has been proposed to address the issue. It hypothesizes that one image can be split into two parts, interest of object to be translated and background to be retained. We extend the assumption that there are several instances belonging to same domain in one image and come up with instance-level image translation. In horse and zebra case, object level image translation always translates all horses or zebras but instance level image translation allows us only translate one of them or desired horses and zebras
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
