NEWS

ICCV 2025 – MIPI Challenge Paper Published, Elevating Hybrid-EVS Deblurring to SOTA

130 teams worldwide competed in the challenge co-organized by AlpsenTek, Pentas and NTU.

The results paper for the ICCV 2025-MIPI “Deblurring for HVS/Hybrid-EVS Camera” challenge – co-organized by AlpsenTek, Pentas and Nanyang Technological University (NTU) – has been published in the ICCV 2025 Workshop proceedings (pp. 3865 – 3875).

Launched in May 2025 and hosted on the Codabench platform, the challenge drew 130 registered teams from industry and academia. Submissions collectively advanced the state of the art in mobile deblurring, lifting image reconstruction quality to SOTA levels.

The paper reviews eight winning solutions: champion NoahTCV, runner-up vivo Camera Research, third-place MiVideoDeblur, plus 404NotFound, IVISLAB, Yunfan Lu, suospring and Attempt++. Every team fused APS image data with EVS event data in its network architecture, exploiting the complementary strengths of the two modalities.

Two trends emerged. First, gains in algorithm performance came with longer inference times, so future MIPI challenges will weigh inference efficiency (latency, model size) alongside image quality (PSNR/SSIM). Second, some teams applied vision-language model (VLM) multimodal fusion frameworks to the task, opening a new route for handling heterogeneous HVS information.

“Breakthroughs in hybrid vision depend on deep coupling of algorithms, hardware and real-world scenarios,” AlpsenTek commented. “We hope the challenge spurs global developer creativity and accelerates industrialization.”

Related News

最近搜索Recent Searches

WeChat Search

AlpsenTek