Research map / 3D vision / Depth and stereo
Improved Efficiency Based on Learned Saccade and Continuous Scene Reconstruction From Foveated Visual Sampling
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a foveated visual sampling system that uses a learned saccade policy and continuous scene reconstruction to improve efficiency of image classification and detection.
Problem
Image classification and detection face a trade-off between accuracy, latency, and energy efficiency, since high-quality images increase computation, reduce camera refresh rates, and increase data communication.
Approach
Uses foveated sampling guided by a learned saccade (gaze) policy, with continuous reconstruction of the scene from the sparse foveated samples, so that full-resolution imaging is avoided while keeping recognition accuracy.
Closest papers by meaning
- Seeing More with Less: Human-like Representations in Vision Models (CVPR 2025 · similarity 0.87)
- Emergence of foveal image sampling from learning to attend in visual scenes (ICLR 2017 · similarity 0.87)
- CanViT: Toward Active-Vision Foundation Models (NeurIPS 2026 · similarity 0.85)
- TransNeXt: Robust Foveal Visual Perception for Vision Transformers (CVPR 2024 · similarity 0.84)
- Finding Visual Saliency in Continuous Spike Stream (AAAI 2024 · similarity 0.84)
- Latent Space Imaging (CVPR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.