Detecting safety helmets is a vital task for industrial management and Internet of Things (IoT) systems. The problem, however, is that deploying high-performance detection models on edge devices is difficult. These devices often struggle with limited resources, leading to high computational redundancy and memory latency. To solve this efficiency issue without losing accuracy, we propose Faster-YOLOv8. This is a lightweight helmet detection algorithm built on Partial Convolution (PConv). Our key move is reconstructing the backbone network: we replaced the original bottleneck layers with a new C2f Faster module. This design cuts down on unnecessary calculations but keeps the feature diversity intact. We ran extensive tests on a self-built safety helmet dataset, and the results prove that our method finds a much better balance between model size and performance. Specifically, Faster-YOLOv8 reduces the parameter count by about 27% (down to 2.19M) and FLOPs by 23% compared to the baseline. At the same time, it maintains a high mAP@0.5 of 87.7%, which is a significant improvement over other lightweight option like GhostNet.
Zhanhe Zhang, Xue Zhang· The 2026 International Confe...· 0 citations
Many flow-based video frame interpolation (VFI) methods synthesize an intermediate frame by estimating optical flow fields, warping the two input frames, and blending the warped observations. These latent flow fields are typically learned through image-level reconstruction supervision without direct flow annotations. In ambiguous regions containing repetitive or stochastic textures, rotating symmetric structures, or fast motion with blur, the matching evidence for a single query may contain multiple comparable and spatially separated peaks. Although the ground-truth intermediate frame provides indirect supervision, it may not uniquely identify the latent correspondence in ambiguous regions.When several locations provide multiple plausible matches, a single-flow estimator can retain only one displacement and discard the remaining candidates. If the selected match is incorrect or inconsistent with those of neighboring pixels, warping samples content from mismatched locations, producing ghosting, structural distortion, or blur.To address this limitation, we propose a multiple hypothesis flow estimation framework that preserves top-K candidate correspondences and selects one per location through a reliability-guided router. Each hypothesis is initialized from a coarse matching anchor and refined separately through anchor-centered local attention. Frame synthesis is thus conditioned on one selected flow-appearance hypothesis rather than a soft combination of candidate motions.Experiments on the proposed MA-HD benchmark and public VFI benchmarks show that our method achieves the best LPIPS and DISTS among the compared methods.
Zibo Su, Jing Kong, Rui-Xing Wang et al.· 0 citations