CVPR 2023poster273 citations

Generalized Decoding for Pixel, Image, and Language

Xueyan Zou, Zi-Yi Dou, Jianwei Yang, Zhe Gan, Linjie Li, Chunyuan Li, Xiyang Dai, Harkirat Behl

Abstract

We present X-Decoder, a generalized decoding model that can predict pixel-level segmentation and language tokens seamlessly. X-Decoder takes as input two types of queries: (i) generic non-semantic queries and (ii) semantic queries induced from text inputs, to decode different pixel-level and token-level outputs in the same semantic space. With such a novel design, X-Decoder is the first work that provides a unified way to support all types of image segmentation and a variety of vision-language (VL) tasks. Further, our design enables seamless interactions across tasks at different granularities and brings mutual benefits by learning a common and rich pixel-level visual-semantic understanding space, without any pseudo-labeling. After pretraining on a mixed set of a limited amount of segmentation data and millions of image-text pairs, X-Decoder exhibits strong transferability to a wide range of downstream tasks in both zero-shot and finetuning settings. Notably, it achieves (1) state-of-the-art results on open-vocabulary segmentation and referring segmentation on eight datasets; (2) better or competitive finetuned performance to other generalist and specialist models on segmentation and VL tasks; and (3) flexibility for efficient finetuning and novel task composition. Code, demo, video and visualization are available at: https://x-decoder-vl.github.io.

BibTeX
@inproceedings{cvpr2023_generalizeddecod,
  title = {Generalized Decoding for Pixel, Image, and Language},
  author = {Xueyan Zou and Zi-Yi Dou and Jianwei Yang and Zhe Gan and Linjie Li and Chunyuan Li and Xiyang Dai and Harkirat Behl and Jianfeng Wang and Lu Yuan and Nanyun Peng and Lijuan Wang and Yong Jae Lee and Jianfeng Gao},
  booktitle = {CVPR 2023},
  year = {2023}
}
Generalized Decoding for Pixel, Image, and Language · CVPR 2023