Minjian Zhang

Ya-nan Guan, Shaonan Zhang, Hang Guo et al.

In this paper, we present a comprehensive overview of the NTIRE 2026 3rd Restore Any Image Model (RAIM) challenge, with a specific focus on Track 3: AI Flash Portrait. Despite significant advancements in deep learning for image restoration, existing models still encounter substantial challenges in real-world low-light portrait scenarios. Specifically, they struggle to achieve an optimal balance among noise suppression, detail preservation, and faithful illumination and color reproduction. To bridge this gap, this challenge aims to establish a novel benchmark for real-world low-light portrait restoration. We comprehensively evaluate the proposed algorithms utilizing a hybrid evaluation system that integrates objective quantitative metrics with rigorous subjective assessment protocols. For this competition, we provide a dataset containing 800 groups of real-captured low-light portrait data. Each group consists of a 1K-resolution low-light input image, a 1K ground truth (GT), and a 1K person mask. This challenge has garnered widespread attention from both academia and industry, attracting over 100 participating teams and receiving more than 3,000 valid submissions. This report details the motivation behind the challenge, the dataset construction process, the evaluation metrics, and the various phases of the competition. The released dataset and baseline code for this track are publicly available from the same \href{https://github.com/zsn1434/AI_Flash-BaseLine/tree/main}{GitHub repository}, and the official challenge webpage is hosted on \href{https://www.codabench.org/competitions/12885/}{CodaBench}.

A. R. Balasubramanian, Mohammad Hossein Khoshechin Jorshari, Rupak Majumdar et al.

We study the estimation problem for concurrent programs: given a bounded program $P$, estimate the number of Mazurkiewicz trace-equivalence classes induced by its interleavings. This quantity informs two practical questions for enumeration-based model checking: how long a model checking run is likely to take, and what fraction of the search space has been covered so far. We first show the counting problem is #P-hard even for restricted programs and, unless $P=NP$, inapproximable within any subexponential factor, ruling out efficient exact or randomized approximation algorithms. We give a Monte Carlo approach to obtain a poly-time unbiased estimator: we convert a stateless optimal DPOR algorithm into an unbiased estimator by viewing its exploration as a bounded-depth, bounded-width tree whose leaves are the maximal Mazurkiewicz traces. A classical estimator by Knuth, when run on this tree, yields an unbiased estimate. To control the variance, we apply stochastic enumeration by maintaining a small population of partial paths per depth whose evolution is coupled. We have implemented our estimator in the JMC model checker and evaluated it on shared-memory benchmarks. With modest budgets, our estimator yields stable estimates, typically within a 20% band, within a few hundred trials, even when the state space has $10^5$--$10^6$ classes. We also show how the same machinery estimates model-checking cost by weighting all explored graphs, not only complete traces. Our algorithms provide the first provable poly-time unbiased estimators for counting traces, a problem of considerable importance when allocating model checking resources.

Hefei Mei, Taijin Zhao, Shiyuan Tang et al.

Few-shot object detection (FSOD) aims to achieve object detection only using a few novel class training data. Most of the existing methods usually adopt a transfer-learning strategy to construct the novel class distribution by transferring the base class knowledge. However, this direct way easily results in confusion between the novel class and other similar categories in the decision space. To address the problem, we propose generating local reverse samples (LRSamples) in Prototype Reference Frames to adaptively adjust the center position and boundary range of the novel class distribution to learn more discriminative novel class samples for FSOD. Firstly, we propose a Center Calibration Variance Augmentation (CCVA) module, which contains the selection rule of LRSamples, the generator of LRSamples, and augmentation on the calibrated distribution centers. Specifically, we design an intra-class feature converter (IFC) as the generator of CCVA to learn the selecting rule. By transferring the knowledge of IFC from the base training to fine-tuning, the IFC generates plentiful novel samples to calibrate the novel class distribution. Moreover, we propose a Feature Density Boundary Optimization (FDBO) module to adaptively adjust the importance of samples depending on their distance from the decision boundary. It can emphasize the importance of the high-density area of the similar class (closer decision boundary area) and reduce the weight of the low-density area of the similar class (farther decision boundary area), thus optimizing a clearer decision boundary for each category. We conduct extensive experiments to demonstrate the effectiveness of our proposed method. Our method achieves consistent improvement on the Pascal VOC and MS COCO datasets based on DeFRCN and MFDC baselines.