InstaFormer: Instance-Aware Image-to-Image Translation with Transformer
This work addresses the problem of enhancing instance-awareness in image translation for computer vision applications, representing an incremental improvement with novel architectural modifications.
The authors tackled instance-aware image-to-image translation by proposing InstaFormer, a Transformer-based network that integrates global and instance-level information using self-attention and AdaIN, achieving improved performance over latest methods as demonstrated in experiments.
We present a novel Transformer-based network architecture for instance-aware image-to-image translation, dubbed InstaFormer, to effectively integrate global- and instance-level information. By considering extracted content features from an image as tokens, our networks discover global consensus of content features by considering context information through a self-attention module in Transformers. By augmenting such tokens with an instance-level feature extracted from the content feature with respect to bounding box information, our framework is capable of learning an interaction between object instances and the global image, thus boosting the instance-awareness. We replace layer normalization (LayerNorm) in standard Transformers with adaptive instance normalization (AdaIN) to enable a multi-modal translation with style codes. In addition, to improve the instance-awareness and translation quality at object regions, we present an instance-level content contrastive loss defined between input and translated image. We conduct experiments to demonstrate the effectiveness of our InstaFormer over the latest methods and provide extensive ablation studies.