arXiv subjects

Jinghao Li

Publications and source records attributed to Jinghao Li.

6 recordsLinked to original sources

arXiv2025

LayerCraft: Enhancing Text-to-Image Generation with CoT Reasoning and Layered Object Integration

Text-to-image (T2I) generation has made remarkable progress, yet existing systems still lack intuitive control over spatial composition, object consistency, and multi-step editing. We present $\textbf{LayerCraft}$, a modular framework that uses large language models (LLMs) as autonomous agents to orchestrate structured, layered image generation and editing. LayerCraft supports two key capabilities: (1) $\textit{structured generation}$ from simple prompts via chain-of-thought (CoT) reasoning, enabling it to decompose scenes, reason about object placement, and guide composition in a controllable, interpretable manner; and (2) $\textit{layered object integration}$, allowing users to insert and customize objects -- such as characters or props -- across diverse images or scenes while preserving identity, context, and style. The system comprises a coordinator agent, the $\textbf{ChainArchitect}$ for CoT-driven layout planning, and the $\textbf{Object Integration Network (OIN)}$ for seamless image editing using off-the-shelf T2I models without retraining. Through applications like batch collage editing and narrative scene generation, LayerCraft empowers non-experts to iteratively design, customize, and refine visual content with minimal manual effort. Code will be released at https://github.com/PeterYYZhang/LayerCraft.

cs.LG↗

arXiv2024

LocalStyleFool: Regional Video Style Transfer Attack Using Segment Anything Model

Previous work has shown that well-crafted adversarial perturbations can threaten the security of video recognition systems. Attackers can invade such models with a low query budget when the perturbations are semantic-invariant, such as StyleFool. Despite the query efficiency, the naturalness of the minutia areas still requires amelioration, since StyleFool leverages style transfer to all pixels in each frame. To close the gap, we propose LocalStyleFool, an improved black-box video adversarial attack that superimposes regional style-transfer-based perturbations on videos. Benefiting from the popularity and scalably usability of Segment Anything Model (SAM), we first extract different regions according to semantic information and then track them through the video stream to maintain the temporal consistency. Then, we add style-transfer-based perturbations to several regions selected based on the associative criterion of transfer-based gradient information and regional area. Perturbation fine adjustment is followed to make stylized videos adversarial. We demonstrate that LocalStyleFool can improve both intra-frame and inter-frame naturalness through a human-assessed survey, while maintaining competitive fooling rate and query efficiency. Successful experiments on the high-resolution dataset also showcase that scrupulous segmentation of SAM helps to improve the scalability of adversarial attacks under high-resolution data.

cs.CV↗

arXiv2023

The Quantum Kalman Decomposition: A Gramian Matrix Approach

The Kalman canonical form for quantum linear systems was derived in \cite{ZGPG18}. The purpose of this paper is to present an alternative derivation by means of a Gramian matrix approach. Controllability and observability Gramian matrices are defined for linear quantum systems, which are used to characterize various subspaces. Based on these characterizations, real orthogonal and block symplectic coordinate transformation matrices are constructed to transform a given quantum linear system to the Kalman canonical form. An example is used to illustrate the main results.

quant-ph↗

arXiv2019

Structural characterization of linear quantum systems with application to back-action evading measurement

The purpose of this paper is to study the structure of quantum linear systems in terms of their Kalman canonical form, which was proposed in a recent paper \cite{ZGPG18}. The spectral structure of quantum linear systems is explored, which indicates that a quantum linear system is both controllable and observable provided that it is Hurwitz stable. A new parameterization method for quantum linear systems is proposed. This parameterization is designed for the Kalman canonical form directly. Consequently, the parameters involved are in a blockwise form in correspondence with the blockwise structure of the Kalman canonical form. This parameter structure can be used to simplify various quantum control design problems. For example, necessary and sufficient conditions for the realization of quantum back-action evading (BAE) measurements are given in terms of these new parameters. Due to their blockwise nature, a small number of parameters are required for realizing quantum BAE measurements.

quant-ph↗

arXiv2018

Purity of Crystalline Strata

Let $p$ be a prime. Let $n\in\mathbb N-\{0\}$. Let $\mathcal C$ be an $F^n$-crystal over a locally noetherian $\mathbb F_p$-scheme $S$. Let $(a,b)\in\mathbb N^2$. We show that the reduced locally closed subscheme of $S$ whose points are exactly those $x\in S$ such that $(a,b)$ is a break point of the Newton polygon of the fiber $\mathcal C_x$ of $\mathcal C$ at $x$ is pure in $S$, i.e., it is an affine $S$-scheme. This result refines and reobtains previous results of de Jong--Oort, Vasiu, and Yang. As an application, we show that for all $m\in \mathbb N$ the reduced locally closed subscheme of $S$ whose points are exactly those $x\in S$ for which the $p$-rank of $\mathcal C_x$ is $m$ is pure in $S$; the case $n=1$ was previously obtained by Deligne (unpublished) and the general case $n\ge 1$ refines and reobtains a result of Zink.

math.AG↗

arXiv2014

Purity Results on $F$-crystals

For an $F^n$-crystal $\mathfrak C$ over a reduced locally Noetherian $\mathbb F_p$-scheme $S$, Vasiu first obtained the purity of the Newton polygon strata of $S$ defined by $\mathfrak C$. Deligne later obtained the purity of the $p$-rank strata of $S$ defined by $\mathfrak C$. We prove a stronger result that for every fixed point $P_0$ in the $xy$-coordinate plane, the reduced locally closed subscheme $S_{P_0}=\{s\in S| NP(\mathfrak C_s)\mbox{ has }P_0\mbox{ as a break point}\}$ of $S$ is pure in $S$ (A weaker result was known for the weak purity of $S_{p_0}$).

math.AG↗