arXiv subjects

Zhongping Zhang

Publications and source records attributed to Zhongping Zhang.

15 recordsLinked to original sources

arXiv2026

Beyond Distortion Robustness: Rethinking Severe Cropping as Erasure-Resilient Message Embedding

Robust message embedding in images is important for multimedia security applications such as copyright protection and content tracing. Existing methods are largely developed under a distortion robustness paradigm, where the embedded signal remains spatially present but is degraded by noise, blur, or compression. Severe cropping poses a fundamentally different challenge because it removes part of the carrier itself, causing partial payload disappearance rather than mere signal corruption. In this paper, we revisit robust message embedding from an erasure-resilience perspective and present CREST, a proof-of-concept framework for severe-cropping-robust embedding. CREST combines coding-theoretic redundancy with neural embedding and recovery by expanding a compact QR message into a redundant spatial payload via LT fountain coding and coupling it with cropping-aware embedding and fragment recovery. Experiments on COCO, DIV2K, and VOC2012 show that CREST improves recovery under severe cropping while maintaining competitive visual quality. Under mixed distortions with an area retention ratio of 0.7, CREST improves TRA from 18.52% to 68.45% and reduces EMR from 13.88% to 4.21% over the strongest baseline. On COCO2017, CREST still achieves 48.55--65.12% TRA when only 30--50% of the image area is retained, whereas all compared baselines fail to recover the message. These results suggest that severe cropping is better understood as an erasure problem rather than a conventional distortion problem, motivating the joint design of neural embedding and coding-based recovery.

cs.CR↗

arXiv2025

Real, Fake, or Manipulated? Detecting Machine-Influenced Text

Large Language Model (LLMs) can be used to write or modify documents, presenting a challenge for understanding the intent behind their use. For example, benign uses may involve using LLM on a human-written document to improve its grammar or to translate it into another language. However, a document entirely produced by a LLM may be more likely to be used to spread misinformation than simple translation (\eg, from use by malicious actors or simply by hallucinating). Prior works in Machine Generated Text (MGT) detection mostly focus on simply identifying whether a document was human or machine written, ignoring these fine-grained uses. In this paper, we introduce a HiErarchical, length-RObust machine-influenced text detector (HERO), which learns to separate text samples of varying lengths from four primary types: human-written, machine-generated, machine-polished, and machine-translated. HERO accomplishes this by combining predictions from length-specialist models that have been trained with Subcategory Guidance. Specifically, for categories that are easily confused (\eg, different source languages), our Subcategory Guidance module encourages separation of the fine-grained categories, boosting performance. Extensive experiments across five LLMs and six domains demonstrate the benefits of our HERO, outperforming the state-of-the-art by 2.5-3 mAP on average.

cs.CL↗

arXiv2024

Machine-Generated Text Localization

Machine-Generated Text (MGT) detection aims to identify a piece of text as machine or human written. Prior work has primarily formulated MGT detection as a binary classification task over an entire document, with limited work exploring cases where only part of a document is machine generated. This paper provides the first in-depth study of MGT that localizes the portions of a document that were machine generated. Thus, if a bad actor were to change a key portion of a news article to spread misinformation, whole document MGT detection may fail since the vast majority is human written, but our approach can succeed due to its granular approach. A key challenge in our MGT localization task is that short spans of text, e.g., a single sentence, provides little information indicating if it is machine generated due to its short length. To address this, we leverage contextual information, where we predict whether multiple sentences are machine or human written at once. This enables our approach to identify changes in style or content to boost performance. A gain of 4-13% mean Average Precision (mAP) over prior work demonstrates the effectiveness of approach on five diverse datasets: GoodNews, VisualNews, WikiText, Essay, and WP. We release our implementation at https://github.com/Zhongping-Zhang/MGT_Localization.

cs.CL↗

arXiv2023

Movie Genre Classification by Language Augmentation and Shot Sampling

Video-based movie genre classification has garnered considerable attention due to its various applications in recommendation systems. Prior work has typically addressed this task by adapting models from traditional video classification tasks, such as action recognition or event detection. However, these models often neglect language elements (e.g., narrations or conversations) present in videos, which can implicitly convey high-level semantics of movie genres, like storylines or background context. Additionally, existing approaches are primarily designed to encode the entire content of the input video, leading to inefficiencies in predicting movie genres. Movie genre prediction may require only a few shots to accurately determine the genres, rendering a comprehensive understanding of the entire video unnecessary. To address these challenges, we propose a Movie genre Classification method based on Language augmentatIon and shot samPling (Movie-CLIP). Movie-CLIP mainly consists of two parts: a language augmentation module to recognize language elements from the input audio, and a shot sampling module to select representative shots from the entire video. We evaluate our method on MovieNet and Condensed Movies datasets, achieving approximate 6-9% improvement in mean Average Precision (mAP) over the baselines. We also generalize Movie-CLIP to the scene boundary detection task, achieving 1.1% improvement in Average Precision (AP) over the state-of-the-art. We release our implementation at github.com/Zhongping-Zhang/Movie-CLIP.

cs.CV↗

arXiv2023

Text-to-image Editing by Image Information Removal

Diffusion models have demonstrated impressive performance in text-guided image generation. Current methods that leverage the knowledge of these models for image editing either fine-tune them using the input image (e.g., Imagic) or incorporate structure information as additional constraints (e.g., ControlNet). However, fine-tuning large-scale diffusion models on a single image can lead to severe overfitting issues and lengthy inference time. Information leakage from pretrained models also make it challenging to preserve image content not related to the text input. Additionally, methods that incorporate structural guidance (e.g., edge maps, semantic maps, keypoints) find retaining attributes like colors and textures difficult. Using the input image as a control could mitigate these issues, but since these models are trained via reconstruction, a model can simply hide information about the original image when encoding it to perfectly reconstruct the image without learning the editing task. To address these challenges, we propose a text-to-image editing model with an Image Information Removal module (IIR) that selectively erases color-related and texture-related information from the original image, allowing us to better preserve the text-irrelevant content and avoid issues arising from information hiding. Our experiments on CUB, Outdoor Scenes, and COCO reports our approach achieves the best editability-fidelity trade-off results. In addition, a user study on COCO shows that our edited images are preferred 35% more often than prior work.

cs.CV↗

arXiv2023

Show, Write, and Retrieve: Entity-aware Article Generation and Retrieval

Article comprehension is an important challenge in natural language processing with many applications such as article generation or image-to-article retrieval. Prior work typically encodes all tokens in articles uniformly using pretrained language models. However, in many applications, such as understanding news stories, these articles are based on real-world events and may reference many named entities that are difficult to accurately recognize and predict by language models. To address this challenge, we propose an ENtity-aware article GeneratIoN and rEtrieval (ENGINE) framework, to explicitly incorporate named entities into language models. ENGINE has two main components: a named-entity extraction module to extract named entities from both metadata and embedded images associated with articles, and an entity-aware mechanism that enhances the model's ability to recognize and predict entity names. We conducted experiments on three public datasets: GoodNews, VisualNews, and WikiText, where our results demonstrate that our model can boost both article generation and article retrieval performance, with a 4-5 perplexity improvement in article generation and a 3-4% boost in recall@1 in article retrieval. We release our implementation at https://github.com/Zhongping-Zhang/ENGINE .

cs.CL↗

arXiv2023

Complex Scene Image Editing by Scene Graph Comprehension

Conditional diffusion models have demonstrated impressive performance on various tasks like text-guided semantic image editing. Prior work requires image regions to be identified manually by human users or use an object detector that only perform well for object-centric manipulations. For example, if an input image contains multiple objects with the same semantic meaning (such as a group of birds), object detectors may struggle to recognize and localize the target object, let alone accurately manipulate it. To address these challenges, we propose a two-stage method for achieving complex scene image editing by Scene Graph Comprehension (SGC-Net). In the first stage, we train a Region of Interest (RoI) prediction network that uses scene graphs and predict the locations of the target objects. Unlike object detection methods based solely on object category, our method can accurately recognize the target object by comprehending the objects and their semantic relationships within a complex scene. The second stage uses a conditional diffusion model to edit the image based on our RoI predictions. We evaluate the effectiveness of our approach on the CLEVR and Visual Genome datasets. We report an 8 point improvement in SSIM on CLEVR and our edited images were preferred by human users by 9-33% over prior work on Visual Genome, validating the effectiveness of our proposed method. Code is available at github.com/Zhongping-Zhang/SGC_Net.

cs.CV↗

arXiv2021

Effectively Leveraging Attributes for Visual Similarity

Measuring similarity between two images often requires performing complex reasoning along different axes (e.g., color, texture, or shape). Insights into what might be important for measuring similarity can can be provided by annotated attributes, but prior work tends to view these annotations as complete, resulting in them using a simplistic approach of predicting attributes on single images, which are, in turn, used to measure similarity. However, it is impractical for a dataset to fully annotate every attribute that may be important. Thus, only representing images based on these incomplete annotations may miss out on key information. To address this issue, we propose the Pairwise Attribute-informed similarity Network (PAN), which breaks similarity learning into capturing similarity conditions and relevance scores from a joint representation of two images. This enables our model to identify that two images contain the same attribute, but can have it deemed irrelevant (e.g., due to fine-grained differences between them) and ignored for measuring similarity between the two images. Notably, while prior methods of using attribute annotations are often unable to outperform prior art, PAN obtains a 4-9% improvement on compatibility prediction between clothing items on Polyvore Outfits, a 5% gain on few shot classification of images using Caltech-UCSD Birds (CUB), and over 1% boost to Recall@1 on In-Shop Clothes Retrieval. Implementation available at https://github.com/samarth4149/PAN

cs.CV↗

arXiv2019

Data-driven Seismic Waveform Inversion: A Study on the Robustness and Generalization

Acoustic- and elastic-waveform inversion is an important and widely used method to reconstruct subsurface velocity image. Waveform inversion is a typical non-linear and ill-posed inverse problem. Existing physics-driven computational methods for solving waveform inversion suffer from the cycle skipping and local minima issues, and not to mention solving waveform inversion is computationally expensive. In recent years, data-driven methods become a promising way to solve the waveform inversion problem. However, most deep learning frameworks suffer from generalization and over-fitting issue. In this paper, we developed a real-time data-driven technique and we call it VelocityGAN, to accurately reconstruct subsurface velocities. Our VelocityGAN is built on a generative adversarial network (GAN) and trained end-to-end to learn a mapping function from the raw seismic waveform data to the velocity image. Different from other encoder-decoder based data-driven seismic waveform inversion approaches, our VelocityGAN learns regularization from data and further impose the regularization to the generator so that inversion accuracy is improved. We further develop a transfer learning strategy based on VelocityGAN to alleviate the generalization issue. A series of experiments are conducted on the synthetic seismic reflection data to evaluate the effectiveness, efficiency, and generalization of VelocityGAN. We not only compare it with existing physics-driven approaches and data-driven frameworks but also conduct several transfer learning experiments. The experiment results show that VelocityGAN achieves state-of-the-art performance among the baselines and can improve the generalization results to some extent.

eess.SP↗

arXiv2019

Earthquake Detection in 1-D Time Series Data with Feature Selection and Dictionary Learning

Earthquakes can be detected by matching spatial patterns or phase properties from 1-D seismic waves. Current earthquake detection methods, such as waveform correlation and template matching, have difficulty detecting anomalous earthquakes that are not similar to other earthquakes. In recent years, machine-learning techniques for earthquake detection have been emerging as a new active research direction. In this paper, we develop a novel earthquake detection method based on dictionary learning. Our detection method first generates rich features via signal processing and statistical methods and further employs feature selection techniques to choose features that carry the most significant information. Based on these selected features, we build a dictionary for classifying earthquake events from non-earthquake events. To evaluate the performance of our dictionary-based detection methods, we test our method on a labquake dataset from Penn State University, which contains 3,357,566 time series data points with a 400 MHz sampling rate. 1,000 earthquake events are manually labeled in total, and the length of these earthquake events varies from 74 to 7151 data points. Through comparison to other detection methods, we show that our feature selection and dictionary learning incorporated earthquake detection method achieves an 80.1% prediction accuracy and outperforms the baseline methods in earthquake detection, including Template Matching (TM) and Support Vector Machine (SVM).

physics.geo-ph↗

arXiv2019

A Data-Driven CO2 Leakage Detection Using Seismic Data and Spatial-Temporal Densely Connected Convolutional Neural Networks

In carbon capture and sequestration, developing effective monitoring methods is needed to detect and respond to CO2 leakage. CO2 leakage detection methods rely on geophysical observations and monitoring sensor network. However, traditional methods usually require the development of site-specific physical models and expert interpretation, and the effectiveness of these methods can be limited to different application locations, operational scenarios, and conditions. In this paper, we developed a novel data-driven leakage detection method based on densely connected convolutional neural networks. Our method differs from conventional leakage monitoring methods by directly learning a mapping relationship between seismic data and the CO2 leakage mass. To account for the spatial and temporal characteristics of seismic data, our novel networks architecture combines 1D and 2D convolutional neural networks. To overcome the computational expense of solving optimization problems, we apply a densely-connecting strategy in our network architecture that reduces the number of network parameters. Based on the features generated by our convolutional neural networks, we further incorporate a long short-term memory network to utilize time-sequential information, which further improves the detection accuracy. Finally, we employ our detection method to synthetic seismic datasets generated based on flow simulations of a hypothetical CO2 storage scenario with injection into a partially compartmentalized sandstone storage reservoir. To evaluate method performance, we conducted multiple experiments including a random leakage test, a sequential test, and a robustness test. Numerical results show that our CO2 leakage detection method successfully detects the leakage and accurately predicts the leakage mass, suggesting that it has the potential for application in monitoring of real CO2 storage sites.

physics.geo-ph↗

arXiv2019

How to Become Instagram Famous: Post Popularity Prediction with Dual-Attention

With a growing number of social apps, people have become increasingly willing to share their everyday photos and events on social media platforms, such as Facebook, Instagram, and WeChat. In social media data mining, post popularity prediction has received much attention from both data scientists and psychologists. Existing research focuses more on exploring the post popularity on a population of users and including comprehensive factors such as temporal information, user connections, number of comments, and so on. However, these frameworks are not suitable for guiding a specific user to make a popular post because the attributes of this user are fixed. Therefore, previous frameworks can only answer the question "whether a post is popular" rather than "how to become famous by popular posts". In this paper, we aim at predicting the popularity of a post for a specific user and mining the patterns behind the popularity. To this end, we first collect data from Instagram. We then design a method to figure out the user environment, representing the content that a specific user is very likely to post. Based on the relevant data, we devise a novel dual-attention model to incorporate image, caption, and user environment. The dual-attention model basically consists of two parts, explicit attention for image-caption pairs and implicit attention for user environment. A hierarchical structure is devised to concatenate the explicit attention part and implicit attention part. We conduct a series of experiments to validate the effectiveness of our model and investigate the factors that can influence the popularity. The classification results show that our model outperforms the baselines, and a statistical analysis identifies what kind of pictures or captions can help the user achieve a relatively high "likes" number.

cs.SI↗

arXiv2018

"Factual" or "Emotional": Stylized Image Captioning with Adaptive Learning and Attention

Generating stylized captions for an image is an emerging topic in image captioning. Given an image as input, it requires the system to generate a caption that has a specific style (e.g., humorous, romantic, positive, and negative) while describing the image content semantically accurately. In this paper, we propose a novel stylized image captioning model that effectively takes both requirements into consideration. To this end, we first devise a new variant of LSTM, named style-factual LSTM, as the building block of our model. It uses two groups of matrices to capture the factual and stylized knowledge, respectively, and automatically learns the word-level weights of the two groups based on previous context. In addition, when we train the model to capture stylized elements, we propose an adaptive learning approach based on a reference factual model, it provides factual knowledge to the model as the model learns from stylized caption labels, and can adaptively compute how much information to supply at each time step. We evaluate our model on two stylized image captioning datasets, which contain humorous/romantic captions and positive/negative captions, respectively. Experiments shows that our proposed model outperforms the state-of-the-art approaches, without using extra ground truth supervision.

cs.CV↗

arXiv2018

Boundary-based Image Forgery Detection by Fast Shallow CNN

Image forgery detection is the task of detecting and localizing forged parts in tampered images. Previous works mostly focus on high resolution images using traces of resampling features, demosaicing features or sharpness of edges. However, a good detection method should also be applicable to low resolution images because compressed or resized images are common these days. To this end, we propose a Shallow Convolutional Neural Network(SCNN), capable of distinguishing the boundaries of forged regions from original edges in low resolution images. SCNN is designed to utilize the information of chroma and saturation. Based on SCNN, two approaches that are named Sliding Windows Detection (SWD) and Fast SCNN, respectively, are developed to detect and localize image forgery region. In this paper, we substantiate that Fast SCNN can detect drastic change of chroma and saturation. In image forgery detection experiments Our model is evaluated on the CASIA 2.0 dataset. The results show that Fast SCNN performs well on low resolution images and achieves significant improvements over the state-of-the-art.

cs.CV↗

arXiv2016

Superconducting nanowire single photon detector at 532 nm and demonstration in satellite laser ranging

Superconducting nanowire single-photon detectors (SNSPDs) at a wavelength of 532 nm were designed and fabricated aiming to satellite laser ranging (SLR) applications. The NbN SNSPDs were fabricated on one-dimensional photonic crystals with a sensitive-area diameter of 42 um. The devices were coupled with multimode fiber (phi=50um) and exhibited a maximum system detection efficiency of 75% at an extremely low dark count rate of <0.1 Hz. An SLR experiment using an SNSPD at a wavelength of 532 nm was successfully demonstrated. The results showed a depth ranging with a precision of ~8.0 mm for the target satellite LARES, which is ~3,000 km away from the ground ranging station at the Sheshan Observatory.

cond-mat.supr-con↗