LiteGfm: A Lightweight Self-supervised Monocular Depth Estimation Framework for Artifacts Reduction via Guided Image Filtering

Zhilin He, Yawei Zhang, Jingchang Mu, Xiaoyue Gu, Tianhao Gu

Abstract: Facing with two significant challenges for monocular depth estimation under a lightweight network, including the preservation of detail information and the artifact reduction of the predicted depth maps, this paper proposes a self-supervised monocular depth estimation framework, called LiteGfm. It contains a DepthNet with an Anti-Artifact Guided (AAG) module and a PoseNet. In the AAG module, a Guided Image Filtering with cross-detail masking is first designed to filter the input features of the decoder for preserving comprehensive detail information. Second, a filter kernel generator is proposed to decompose the Sobel operator along the vertical and horizontal axes for achieving cross-detail masking, which better captures the structure and edge feature for minimizing artifacts. Furthermore, a boundary-aware loss between the reconstructed and input images is presented to preserve high-frequency details for decreasing artifacts. Extensive experimental results demonstrate that LiteGfm under 1.9M parameters gets more optimal performance than state-of-the-art methods.

Primary Subject Area: [Content] Media Interpretation

Relevance To Conference: Scene understanding plays a key role in multimedia: extracting content, retrieving and classifying, and intelligent editing and generation. It recognizes objects and scene actions in images and videos, making content processing more intelligent and efficient. Our paper is to tackle the depth estimation task in scene understanding. Given that many applications in robotics, autonomous driving, and augmented reality depend on depth maps, which convey the 3D geometry of a scene, this greatly facilitates the advancement of artificial intelligence. Simultaneously, it provides new guidance for the visual domain of multimedia, positively impacting subsequent multimedia tasks involving vision, such as multimodal fusion.

Submission Number: 4367