MCL Research on Green Image Steganalysis

Steganography for spatial images is a technique that involves hiding secret information within digital images in a way that is imperceptible to the human eye. It considers the characteristics and features of the cover media in a local region, and embed the secret information in a manner that is visually or statistically inconspicuous.

As the other side of the coin, steganalysis, is the way of detecting the embedded image (often we call it stego image). There has been numerous works in detecting the hidden information. Traditionally, people use heuristic features and ensemble of machine learning models for detection. It soon becomes feeble in detecting content adaptive steganographic images. After the emerging of neural networks, researchers start to use deep models to detect the weak stego signals by extracting features and doing classification in a whole. Different from the traditional steganalysis and deep learning based steganalysis methods, we propose a novel steganalysis scheme, which is a green steganalysis method.

We first scatter the whole image into small blocks, and then perform anomaly detection on block levels. This step will give us an indication of the likelihood such that this block is embedded or not. Next, we train an embed location detector, to help us locate the blocks that are more discriminant than others. Finally, blocks that are selected from previous step will be fused together and make image-level decision by ensemble classifier. Our architecture is completely explainable, and computationally efficient.

— by Yao Zhu

By |April 22nd, 2023|News|Comments Off on MCL Research on Green Image Steganalysis|
  • Permalink Gallery

    MCL Genealogical Ancestry Series – Abraham Gotthelf Kastner

MCL Genealogical Ancestry Series – Abraham Gotthelf Kastner

Abraham Gotthelf Kastner (27 September 1719 – 20 June 1800) was a German mathematician and epigrammatist.

Kästner was the son of law professor Abraham Kästner. Starting from 1731 (aged 12), he studied law, philosophy, physics, mathematics and metaphysics in Leipzig. He was appointed a Notary in 1733, when he was 14. He gained his habilitation (PhD degree in Europe) from the University of Leipzig in 1739 at age 20.  In his early careers, he lectured mathematics, philosophy, logic and law in University of Leipzig after his habilitation. He soon became an associate professor in 1746, at age 27. In 1751 he was elected a member of the Royal Swedish Academy of Sciences, which have famous fellows such as Isaac Newton, Charles Darwin, Alan Turing, Stephen Hawking, etc. In 1756, at his 37, he took up a position as full professor of natural philosophy and geometry at the University of Göttingen. His notable doctoral students include Johann Pfaff, who is the doctoral advisor of Carl Friedrich Gauss). Kästner died in 1800 in Göttingen, at age 81.

Kästner has numerous mathematical writings, including Anfangsgründe der Mathematik (“Foundations of Mathematics”) (Göttingen 1758-69, 4 volumes; 6th edition 1800) and Geschichte der Mathematik (“History of Mathematics”) (Göttingen 1796-1800, 4 volumes). Geschichte der Mathematik is considered an astute work, but lacks a comprehensive overview of all subsections of mathematics. Besides his contribution in math, he is more well-known for his poems, which were notable for their biting humour and sharp irony on different contemporary personalities.

As his descendants, we know Kästner as a talented scholar and devoted researcher in mathematics, philosophy, logic and law. His rich contributions shall be remembered the same as himself.

By |April 16th, 2023|News|Comments Off on MCL Genealogical Ancestry Series – Abraham Gotthelf Kastner|
  • Permalink Gallery

    Professor Kuo Delivered Keynote Speech in A Workshop Held by IEEE Hong Kong Section

Professor Kuo Delivered Keynote Speech in A Workshop Held by IEEE Hong Kong Section

Professor C.-C. Jay Kuo, Director of MCL, was invited by the IEEE Hong Kong Chapter to deliver a keynote speech titled “On the 2nd AI wave: toward interpretable, reliable, and sustainable AI” in the 2-Day IEEE Workshop on Deep Learning held at the Hong Kong Science Park, Pak Shek Kok, New Territories, Hong Kong, on April 3, 2023. In his keynote, Professor Kuo pointed out some concerns with the deep learning methodology commonly employed by AI researchers and presented a new green learning paradigm as an alternative that can achieve comparable performance at a lower computational cost.
During his visit to Hong Kong from April 2-7, Professor Kuo visited several universities and gave the following talks:

“Green learning models for point cloud analysis,” Chinese University of Hong Kong.
“Green learning models for point cloud analysis,” Hong Kong University of Science and Technology.
“A lightweight blind image quality assessment (BIQA) method,” Hong Kong Polytechnic University.
“A lightweight blind image quality assessment (BIQA) method,” City University of Hong Kong.

He used “point cloud analysis” and “blind image quality assessment” as examples, explained how to use the green learning methodology to solve them, and compared the pros and cons of deep learning and green learning.
Furthermore, Professor Kuo gave a special seminar on “How to conduct high-quality research and manage large research groups” to faculty members at the Caritas Institute of Higher Education.

By |April 9th, 2023|News|Comments Off on Professor Kuo Delivered Keynote Speech in A Workshop Held by IEEE Hong Kong Section|

Congratulations to Pranav Kadam for Passing His Defense

Congratulations to Pranav Kadam for passing his defense on Mar. 22. Pranav’s thesis is titled “Green Learning for 3D Point Cloud Data Processing”. His Dissertation Committee includes Jay Kuo (Chair), Antonio Ortega, and Aiichiro Nakano (Outside Member). Here we invite Pranav to share about his PhD thesis and his PhD experience.


Thesis Abstract: 
3D Point Cloud processing and analysis has attracted a lot of attention in present times due to the numerous applications such as in autonomous driving, computer graphics, and robotics. In this dissertation, we focus on the problems of point cloud registration, pose estimation, rotation invariant classification, odometry and scene flow estimation. These tasks are important in the realization of a 3D vision system. Rigid registration aims at finding a 3D transformation consisting of rotation and translation that optimally aligns two point clouds. The next two tasks focus on object-level analysis. For pose estimation, we predict the 6-DOF pose of an object with respect to a chosen frame of reference. Rotation invariant classification aims at classifying 3D objects which are arbitrarily rotated. The latter two problems are for outdoor environments. In odometry, we want to estimate the incremental motion of an object using the point cloud scans captured by it at every instance. While the scene flow estimation task aims at determining the point-wise flow between two consecutive point clouds.
3D perception using point clouds is dominated by deep learning methods nowadays. However, large scale learning on point clouds with deep learning techniques has several issues which are often overlooked. This research is based on the green learning (GL) paradigm and focuses on interpretability, smaller training times and smaller model size. Using GL, we separate the feature learning process from the decision. Features are [...]

By |April 6th, 2023|News|Comments Off on Congratulations to Pranav Kadam for Passing His Defense|

MCL Research on Point-Cloud-based 3D Scene Flow Estimation

3D scene flow aims at finding the point-wise 3D displacement between consecutive point cloud scans. It finds applications in areas such as dynamic scene segmentation and may also guide inter-prediction in compression of dynamically acquired point clouds. We propose a green and interpretable 3D scene flow estimation method for the autonomous driving scenario and name it “PointFlowHop” [1]. We decompose our solution into vehicle ego-motion and object motion components.
The vehicle ego-motion is first compensated using the GreenPCO method which was recently proposed for the task of point cloud odometry estimation. Then, we divide the scene points into two classes – static and moving. The static points do not have any motion and can be assigned only the ego-motion component. The motion of the moving points is analyzed later. For classification, we use a lightweight XGBoost classifier with a 5-dimensional shape and motion feature as the input. Later, moving points are grouped into moving objects using DBSCAN clustering algorithm. Furthermore, the moving objects from the two point clouds are associated using the nearest centroids algorithm. An additional refinement step ensures reclassification of previously misclassified moving points. A rigid flow model is established for each object. Finally, the flow in local regions is refined assuming local scene rigidity.
PointFlowHop method adopts the green learning (GL) paradigm. The task-agnostic nature of the feature learning process in GL enables scene flow estimation through seamless modification and extension of prior related GL methods like R-PointHop and GreenPCO. Furthermore, a large number of operations in PointFlowHop are not performed during training. The ego-motion and object-level motion is optimized in inference only. Similarly, the moving points are grouped into objects only during inference. This makes the training process much faster [...]

By |March 28th, 2023|News, Research|Comments Off on MCL Research on Point-Cloud-based 3D Scene Flow Estimation|
  • Permalink Gallery

    MCL Research on Mask-Guided Image Synthesis Presented at AAAI-23

MCL Research on Mask-Guided Image Synthesis Presented at AAAI-23

Dr. Rouhsedaghat, a MCL alumna graduated last Summer, recently presented a work[1] on image synthesis related to her PhD thesis in AAAI-23. Here is the presentation summary from Dr. Rouhsedaghat:

We offer a method for one-shot mask-guided image synthesis that allows controlling manipulations of a single image by inverting a quasi-robust classifier equipped with strong regularizers. Our proposed method, entitled MAGIC , leverages structured gradients from a pre-trained quasi-robust classifier to better preserve the input semantics while preserving its classification accuracy, thereby guaranteeing credibility in the synthesis. Unlike current methods that use complex primitives to supervise the process or use attention maps as a weak supervisory signal, MAGIC aggregates gradients over the input, driven by a guide binary mask that enforces a strong, spatial prior. MAGIC implements a series of manipulations with a single framework achieving shape and location control, intense non-rigid shape deformations, and copy/move operations in the presence of repeating objects and gives users firm control over the synthesis by requiring to simply specify binary guide masks. Our study and findings are supported by various qualitative comparisons with the state-of-the-art on the same images sampled from ImageNet and quantitative analysis using machine perception along with a user survey of 100+ participants that endorse our synthesis quality.

[1]Rouhsedaghat, Mozhdeh, et al. “MAGIC: Mask-Guided Image Synthesis by Inverting a Quasi-Robust Classifier.” arXiv preprint arXiv:2209.11549 (2022).

By |March 5th, 2023|News, Research|Comments Off on MCL Research on Mask-Guided Image Synthesis Presented at AAAI-23|

MCL Research on Sentence Similarity Modeling

Sentence similarity evaluation has a wide range of applications in natural language processing, such as semantic similarity computation, text generation evaluation, and information retrieval. As one of the word-alignment-based methods, Word Mover’s Distance (WMD)[1] formulates text similarity evaluation as a minimum-cost flow problem. It finds the most efficient way to align the information between text sequences through a flow network defined by word-level similarities. By assigning flows to individual words, WMD computes text dissimilarity as the minimum cost of moving words’ flows from one sentence to another based on pre-trained word embeddings.

However, a naive WMD method does not perform well on sentence similarity evaluation for several reasons.
– First, WMD assigns word flow based on words’ frequency in a sentence. This frequency-based word weighting scheme is weak in capturing word importance when considering the statistics of the whole corpus.
– Second, the distance between words solely depends on the embedding of isolated words without considering the contextual and structural information of input sentences. Since the meaning of a sentence depends on individual words as well as their interaction, simply considering the alignment between individual words is deficient in evaluating sentence similarity.

MCL proposed a new syntax-aware word flow calculation method, Syntax-aware Word Mover’s Distance (SynWMD)[2], for sentence similarity evaluation.
– Words are first represented as a weighted graph based on the co-occurrence statistics obtained by dependency parsing trees. Then, a PageRank-based algorithm is used to infer word importance.
– The word distance model in WMD is enhanced by the context extracted from dependency parse trees, which is illustrated in Figure 1. The contextual information of words and structural information of sentences are explicitly modeled as additional subtree embeddings.
– As shown in Table 1, we [...]

By |September 4th, 2022|News|Comments Off on MCL Research on Sentence Similarity Modeling|

MCL Research on Green Blind Image Quality Assessment

Image quality assessment (IQA) aims to evaluate image quality at various stages of image processing such as image acquisition, transmission, and compression. Based on the availability of undistorted reference images, objective IQA can be classified into three categories [1]: full-reference (FR), reduced-referenced (RR) and no-reference (NR). The last one is also known as blind IQA (BIQA). FR-IQA metrics have achieved high consistency with human subjective evaluation. Many FR-IQA methods have been well developed in the last two decades such as SSIM [2] and FSIM [3]. RR-IQA metrics only utilize features of reference images for quality evaluation. In some application scenarios (e.g., image receivers), users cannot access reference images so that NR-IQA is the only choice. BIQA methods attract growing attention in recent years.

Generally speaking, conventional BIQA methods consist of two stages: 1) extraction of quality-aware features and 2) adoption of a regression model for quality score prediction. As the amount of user generated images grows rapidly, a handcrafted feature extraction method is limited in its power of modeling a wide range of image content and distortion characteristics. Deep neural networks (DNNs) achieve great success in blind image quality assessment (BIQA) with large pre-trained models in recent years. However, their solutions cannot be easily deployed at mobile or edge devices, and a lightweight solution is desired.

In this work, we propose a novel BIQA model, called GreenBIQA, that aims at high performance, low computational complexity and a small model size. GreenBIQA adopts an unsupervised feature generation method and a supervised feature selection method to extract quality-aware features. Then, it trains an XGBoost regressor to predict quality scores of test images. We conduct experiments on four popular IQA datasets, which include two synthetic-distortion and two authentic-distortion [...]

By |August 30th, 2022|News|Comments Off on MCL Research on Green Blind Image Quality Assessment|

MCL Research on Effective Knowledge Graph Embedding

Knowledge Graph encodes human-readable information and knowledge in graph format. Triples, denoted by (h,r,t), are basic elements of a KG, where h and t are head and tail entities while r is the relation connecting them. Both manual effort by domain experts and automated information extraction algorithms have contributed to the creation of many existing Knowledge Graphs today. However, given the limited information accessible to each individual and the limitation of algorithms, it is nearly impossible for a Knowledge Graph to perfectly capture every single piece of facts about the world. As such, Knowledge Graphs are often incomplete and many researchers have developed different algorithms to predict missing facts in Knowledge Graphs. Knowledge Graph Embedding models were first proposed to mainly solve the Knowledge Graph Completion problem. Besides, embedding models can also be useful in solving many downstream tasks such as entity classification and entity alignment.

MCL has been recently working on Effective Knowledge Graph Embedding. Translation, rotation, and scaling are three commonly used geometric manipulation operations in image processing. Besides, some of them are successfully used in developing effective knowledge graph embedding (KGE) models such as TransE and RotatE. Inspired by the synergy, we propose a new KGE model by leveraging all three operations. Since translation, rotation, and scaling operations are cascaded to form a compound one, the new model is named CompoundE. By casting CompoundE in the framework of group theory, we show that quite a few distanced-based KGE models are special cases of CompoundE. CompoundE extends the simple distance-based scoring functions to relation-dependent compound operations on head and/or tail entities. To demonstrate the effectiveness of CompoundE, we conduct experiments on three popular knowledge graph completion datasets. Experimental results show that CompoundE consistently [...]

By |August 22nd, 2022|News|Comments Off on MCL Research on Effective Knowledge Graph Embedding|

MCL Research on Semi-Supervised Feature Learning

Traditional machine learning algorithms are susceptible to the curse of feature dimensionality [1]. Their computational complexity increases with high dimensional features. Redundant features may not be helpful in discriminating classes or reducing regression error, and they should be removed. Sometimes, redundant features may even produce negative effects as their number grows. Their detrimental impact should be minimized or controlled. To deal with these problems, feature learning techniques using feature selection are commonly applied as a data pre-processing step or part of the data analysis to simplify the complexity of the model. Feature selection techniques involve the identification of a subspace of discriminant features from the input, which describe the input data efficiently, reduce effects from noise or irrelevant features, and provide good prediction results.
Inspired by information theory and the decision tree, a novel supervised feature selection methodology is proposed recently in MCL. The resulting tests are called the discriminant feature test (DFT) and the relevant feature test (RFT) for classification and regression tasks, respectively [2]. The proposed methods belong to the filter methods of feature selection, which give a score to each dimension and select features based on feature ranking. The scores are measured by the weighted entropy and the weighted MSE for DFT and RFT, which reflect the discriminant power and relevance degree to classification and regression targets, respectively. It is shown by experimental results that DFT and RFT can select a lower dimensional feature subspace distinctly and robustly while maintaining high decision performance.
The proposed methods work well in the semi-supervised scenario, where useful feature set learnt in a limited number of labeled data has high intersection over union (IoU) compared to giving the full set of labeled training data. Examples [...]

By |August 15th, 2022|News|Comments Off on MCL Research on Semi-Supervised Feature Learning|