{"id":34,"date":"2018-10-30T14:08:36","date_gmt":"2018-10-30T14:08:36","guid":{"rendered":"https:\/\/cchsu.info\/wordpress\/?page_id=34"},"modified":"2026-08-09T18:12:49","modified_gmt":"2026-08-09T10:12:49","slug":"research","status":"publish","type":"page","link":"https:\/\/cchsu.info\/wordpress\/research\/","title":{"rendered":"Research"},"content":{"rendered":"
\n
\n\t

Language:<\/strong> English | \u7e41\u9ad4\u4e2d\u6587<\/a><\/p>\n

Research Vision \u7814\u7a76\u4e3b\u8ef8<\/h1>\n

Advanced Computer Vision Lab \u2014 A<\/strong>ssured C<\/strong>omputer V<\/strong>ision: L<\/strong>ean, A<\/strong>utonomous, B<\/strong>road-Spectrum<\/p>\n

As generative AI blurs the boundary between authentic and fabricated media, autonomous systems demand vision that never fails silently, and Earth observation enters a data-rich new era, the bar for deployable visual intelligence keeps rising. ACVLab responds with four interlocking research pillars.<\/p>\n

Assured Visual Intelligence<\/b> ensures that every visual AI output can be trusted \u2014 whether detecting DeepFakes under heavy compression, defending against adversarial perturbations, or authenticating media through proactive watermarking \u2014 providing the accountability that forensic, medical, and regulatory settings require.<\/p>\n

Efficient Computing Systems<\/b> rethink computation at every level of abstraction: prefix-scan reformulations of exact attention (ELSA), bitstream-level forensics that skip pixel decoding entirely, adaptive quantization that preserves accuracy at ultra-low bit widths (QuantTune\/FracQuant), and joint transmission-restoration for bandwidth-constrained satellites \u2014 cutting latency, memory, and energy cost for sustainable, real-time deployment.<\/p>\n

Autonomous Visual Perception<\/b> extends vision from 2D images into 3D physical space: material-aware scene reconstruction with hyperspectral unmixing, BEV adversarial defense for self-driving (BFDM), physics-aligned shadow and reflection removal that feeds robust features to downstream robotic pipelines (PhaSR, ReflexSplit), and uncertainty-aware 3D annotation for autonomous driving datasets.<\/p>\n

Broad-Spectrum Scientific Sensing<\/b> pushes perception beyond the visible: universal hyperspectral restoration via vision-language prompts (PromptHSI), real-time CubeSat compressed sensing recognized with the Future Technology Award, hyperspectral pansharpening through sparse spectral representations (S3<\/sup>RNet), and cross-spectral forgery detection that reveals manipulation invisible to RGB analysis.<\/p>\n

These pillars do not operate in isolation. Hyperspectral forensics merges trust with spectral sensing. On-satellite real-time inference merges efficiency with broad-spectrum data. BEV adversarial defense merges trust with embodied perception. This cross-pillar synergy is not accidental \u2014 it reflects a single underlying conviction: deployment-grade visual intelligence must be simultaneously trustworthy, efficient, embodied, and perceptually complete.<\/p>\n

Adjacent computational direction.<\/strong> Beyond the four visual pillars, we also explore lean computational systems for emerging workloads such as resource-aware quantum simulation. This is an extension of the lab's deployment-oriented systems perspective, not a fifth visual pillar.<\/p>\n

\n
Research view<\/strong>All cards remain visible; the selected option changes priority\/order only. Within each priority group, cards remain newest first.<\/span><\/div>\n