Abysstree渊林ovc

Projects项目

Research & Project Portfolio研究与项目作品集

Research and portfolio work around multimodal fusion perception, computational imaging, AI camera restoration, and deployable visual systems.围绕多模态融合感知、计算成像、AI 相机图像恢复和可部署视觉系统展开的研究与作品集工作。

Paper line论文方向

RobustIRFusion

Infrared-visible fusion / reliability / misalignment红外-可见光融合 / 可靠性 / 错位

The active paper line. It frames fusion as a robustness problem: how should a system behave when infrared and visible inputs are not perfectly aligned or equally reliable?当前活跃的论文方向。它把融合视为一个鲁棒性问题:当红外和可见光输入并不完全对齐、也并不同等可靠时,系统应该如何表现?

Demo lineDemo 方向

Photo Moire Removal

AI camera restoration / demoireing / before-after demoAI 相机图像恢复 / 去摩尔纹 / 前后对比 Demo

A compact portfolio demo under the AI camera restoration line. The first milestone is a trustworthy baseline gallery and local before-after experience.AI 相机图像恢复方向下的一个紧凑作品集 Demo。第一阶段目标是做出可信的基线结果画廊和本地前后对比体验。

Long-term长期方向

AI Camera Restoration

Computational imaging / ISP-aware restoration / deployable systems计算成像 / ISP 感知恢复 / 可部署系统

The broader portfolio direction connecting optoelectronics, computational imaging, ISP-aware enhancement, and practical visual restoration systems.更长期的作品集方向,用来连接光电背景、计算成像、ISP 感知增强和实际可用的视觉恢复系统。