Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces JEPADepth, a self-supervised monocular depth estimation framework that enhances traditional photometric reconstruction losses with a masked prediction loss derived from Image Joint-Embedding Predictive Architectures (I-JEPA). By leveraging a pretrained DINOv3 Vision Transformer encoder, the method enables a predictor to infer target-region embeddings from visible context-region embeddings, significantly improving depth estimation performance without altering the inference-time architecture. Experimental results on the KITTI dataset demonstrate that JEPADepth outperforms both state-of-the-art transformer-based methods and strong CNN baselines, achieving top performance in zero-shot transfer scenarios on Make3D and Cityscapes.
JEPADepth achieves superior monocular depth estimation by integrating masked predictive learning, outperforming traditional methods without increasing deployment costs.
Self-supervised monocular depth estimation typically relies on photometric reconstruction losses that couple depth, pose, and appearance assumptions. In this paper, we propose JEPADepth, a self-supervised monocular depth framework that incorporates a complementary training objective inspired by Image Joint-Embedding Predictive Architectures (I-JEPA) for self-supervised depth learning. Our method augments a standard photometric pipeline with a masked prediction loss computed in the representation space of a pretrained DINOv3 Vision Transformer encoder. A predictor infers target-region embeddings from visible context-region embeddings under structured masking, and is discarded along with the target encoder at inference time, adding no deployment cost. On KITTI, adding the JEPA objective consistently improves performance over the same DINOv3-based photometric baseline, without changing the inference-time architecture. Compared to prior monocular self-supervised methods, JEPADepth is competitive with state-of-the-art transformer-based approaches and outperforms strong CNN-based baselines on the standard benchmark. In zero-shot transfer (trained on KITTI and evaluated without fine-tuning), JEPADepth achieves the best or near-best performance among the compared methods on both Make3D and Cityscapes across multiple metrics.