ImageBind: One Embedding Space To Bind Them All
We present ImageBind, an approach to learn a joint embedding across six different modalities - images, text, audio, depth, thermal, and IMU data. We show that all combinations of paired data are not necessary to train such a joint embedding, and only image-paired data is sufficient to bind the modalities together. ImageBind can leverage recent large scale vision-language models, and extends their zero-shot capabilities to new modalities just by using their natural pairing with images. It enables novel emergent applications 'out-of-the-box' including cross-modal retrieval, composing modalities with arithmetic, cross-modal detection and generation. The emergent capabilities improve with the strength of the image encoder and we set a new state-of-the-art on emergent zero-shot recognition tasks across modalities, outperforming specialist supervised models. Finally, we show strong few-shot recognition results outperforming prior work, and that ImageBind serves as a new way to evaluate vision models for visual and non-visual tasks.
PDF Abstract CVPR 2023 PDF CVPR 2023 AbstractCode
Tasks
Cross-Modal Retrieval
Multimodal Deep Learning
Retrieval
Sound Prompted Semantic Segmentation
Speech Prompted Semantic Segmentation
Temporal Relation Extraction
Zero-shot Audio Classification
Zero-shot Classification (unified classes)
Zero-Shot Environment Sound Classification
Zero-Shot Learning
Zero-shot Scene Classification (unified classes)
Zero-shot Text to Audio Retrieval
Zero-Shot Video Retrieval
ADE20K
NYUv2
AudioSet
MSR-VTT
SUN RGB-D
ESC-50
AudioCaps
Clotho
VGG-Sound
LLVIP
