arXiv subjects

Bowen Lin

Publications and source records attributed to Bowen Lin.

7 recordsLinked to original sources

arXiv2026

TIDE: Text-Informed Dynamic Extrapolation with Step-Aware Temperature Control for Diffusion Transformers

Diffusion Transformer (DiT) faces challenges when generating images with higher resolution compared at training resolution, causing especially structural degradation due to attention dilution. Previous approaches attempt to mitigate this by sharpening attention distributions, but fail to preserve fine-grained semantic details and introduce obvious artifacts. In this work, we analyze the characteristics of DiTs and propose TIDE, a training-free text-to-image (T2I) extrapolation method that enables generation with arbitrary resolution and aspect ratio without additional sampling overhead. We identify the core factor for prompt information loss, and introduce a text anchoring mechanism to correct the imbalance between text and image tokens. To further eliminate artifacts, we design a dynamic temperature control mechanism that leverages the pattern of spectral progression in the diffusion process. Extensive evaluations demonstrate that TIDE delivers high-quality resolution extrapolation capability and integrates seamlessly with existing state-of-the-art methods.

cs.CV↗

arXiv2026

Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams

Real-time understanding of long video streams remains challenging for multimodal large language models (VLMs) due to redundant frame processing and rapid forgetting of past context. Existing streaming systems rely on fixed-interval decoding or cache pruning, which either produce repetitive outputs or discard crucial temporal information. We introduce Event-VStream, an event-aware framework that represents continuous video as a sequence of discrete, semantically coherent events. Our system detects meaningful state transitions by integrating motion, semantic, and predictive cues, and triggers language generation only at those boundaries. Each event embedding is consolidated into a persistent memory bank, enabling long-horizon reasoning while maintaining low latency. Across OVOBench-Realtime, and long-form Ego4D evaluations, Event-VStream achieves competitive performance. It improves over a VideoLLM-Online-8B baseline by +10.4 points on OVOBench-Realtime, achieves performance close to Flash-VStream-7B despite using only a general-purpose LLaMA-3-8B text backbone, and maintains around 70% GPT-5 win rate on 2-hour Ego4D streams.

cs.CV↗

arXiv2026

SDiT: Semantic Region-Adaptive for Diffusion Transformers

Diffusion Transformers (DiTs) achieve state-of-the-art performance in text-to-image synthesis but remain computationally expensive due to the iterative nature of denoising and the quadratic cost of global attention. In this work, we observe that denoising dynamics are spatially non-uniform-background regions converge rapidly while edges and textured areas evolve much more actively. Building on this insight, we propose SDiT, a Semantic Region-Adaptive Diffusion Transformer that allocates computation according to regional complexity. SDiT introduces a training-free framework combining (1) semantic-aware clustering via fast Quickshift-based segmentation, (2) complexity-driven regional scheduling to selectively update informative areas, and (3) boundary-aware refinement to maintain spatial coherence. Without any model retraining or architectural modification, SDiT achieves up to 3.0x acceleration while preserving nearly identical perceptual and semantic quality to full-attention inference.

cs.CV↗

arXiv2025

Diverse Normal Prototypes-Guided Contrastive Reconstruction for Medical Anomaly Detection

Anomaly detection in medical images is challenging due to limited annotations and the domain gap. Existing reconstruction-based methods often rely on frozen pre-trained encoders, restricting adaptation to domain-specific patterns and degrading localization accuracy. Meanwhile, prototype-based learning offers interpretable representations but commonly suffers from prototype collapse, where a few prototypes dominate training and reduce diversity. To address these issues, we propose DNP-ConFormer, a unified framework that integrates a trainable encoder with prototype-guided reconstruction and a Diversity-Aware Alignment Loss. A momentum encoder enables stable domain-adaptive representation learning, while a lightweight Prototype Extractor discovers informative normal prototypes and injects them into the decoder via attention to guide reconstruction. The proposed alignment objective further encourages balanced feature-to-prototype assignments, effectively mitigating prototype collapse. Extensive experiments on multiple medical imaging benchmarks demonstrate improved representation quality and anomaly localization compared with prior methods. Visualization and prototype assignment analyses further validate the effectiveness and interpretability of our approach. The code is available at https://github.com/liluhu0/DNP-ConFormer.

cs.CV↗

arXiv2025

An energy optimization method based on mixed-integer model and variational quantum computing algorithm for faster IMPT

Intensity-modulated proton therapy (IMPT) offers superior dose conformity with reduced exposure to surrounding healthy tissues compared to conventional photon therapy. Improving IMPT delivery efficiency reduces motion-related uncertainties, enhances plan robustness, and benefits breath-hold techniques by shortening treatment time. Among various factors, energy switching time plays a critical role, making energy layer optimization (ELO) essential. This work develops an energy layer optimization method based on mixed integer model and variational quantum computing algorithm to enhance the efficiency of IMPT. The energy layer optimization problem is modeled as a mixed-integer program, where continuous variables optimize the dose distribution and binary variables indicate energy layer selection. To solve it, iterative convex relaxation decouples the dose-volume constraints, followed by the alternating direction method of multipliers (ADMM) to separate mixed-variable optimization and the minimum monitor unit (MMU) constraint. The resulting beam intensity subproblem, subject to MMU, either admits a closed-form solution or is efficiently solvable via conjugate gradient. The binary subproblem is cast as a quadratic unconstrained binary optimization (QUBO) problem, solvable using variational quantum computing algorithms. With nearly the same plan quality, the proposed method noticeable reduces the number of the used energies. For example, compared to conventional IMPT, QC can reduce the number of energy layers from 61 to 35 in HN case, from 56 to 35 in lung case, and from 59 to 32 to abdomen case. The reduced number of energies also results in fewer delivery time, e.g., the delivery time is reduced from 100.6, 232.0, 185.3 seconds to 90.7, 215.4, 154.0 seconds, respectively.

physics.med-ph↗

arXiv2019

Fast LLMMSE filter for low-dose CT imaging

Low-dose X-ray CT technology is one of important directions of current research and development of medical imaging equipment. A fast algorithm of blockwise sinogram filtering is presented for realtime low-dose CT imaging. A nonstationary Gaussian noise model of low-dose sinogram data is proposed in the low-mA (tube current) CT protocol. Then, according to the linear minimum mean square error principle, an adaptive blockwise algorithm is built to filter contaminated sinogram data caused by photon starvation. A moving sum technique is used to speed the algorithm into a linear time one, regardless of the block size and thedata range. The proposedfast filtering givesa better performance in noise reduction and detail preservation in the reconstructed images,which is verified in experiments on simulated and real data compared with some related filtering methods.

cs.CV↗

arXiv2019

Optical Fringe Patterns Filtering Based on Multi-Stage Convolution Neural Network

Optical fringe patterns are often contaminated by speckle noise, making it difficult to accurately and robustly extract their phase fields. To deal with this problem, we propose a filtering method based on deep learning, called optical fringe patterns denoising convolutional neural network (FPD-CNN), for directly removing speckle from the input noisy fringe patterns. Regularization technology is integrated into the design of deep architecture. Specifically, the FPD-CNN method is divided into multiple stages, each stage consists of a set of convolutional layers along with batch normalization and leaky rectified linear unit (Leaky ReLU) activation function. The end-to-end joint training is carried out using the Euclidean loss. Extensive experiments on simulated and experimental optical fringe patterns,especially finer ones with high-density regions, show that the proposed method is competitive with some state-of-the-art denoising techniques in spatial or transform domains, efficiently preserving main features of fringe at a fairly fast speed.

cs.CV↗