Controllable image generation based on causal representation learning

Artificial intelligence generated content(AIGC)has emerged as an indispensable tool for producing large-scale content in various forms,such as images,thanks to the significant role that AI plays in imitation and production.However,interpretability and controllability remain challenges.Existing AI methods often face challenges in producing images that are both flexible and controllable while considering causal relationships within the images.To address this issue,we have developed a novel method for causal controllable image generation(CCIG)that combines causal representation learning with bi-directional generative adversarial networks(GANs).This approach enables humans to control image attributes while considering the rationality and interpretability of the generated images and also allows for the generation of counterfactual images.The key of our approach,CCIG,lies in the use of a causal structure learning module to learn the causal relationships between image attributes and joint optimization with the encoder,generator,and joint discriminator in the image generation module.By doing so,we can learn causal representations in image’s latent space and use causal intervention operations to control image generation.We conduct extensive experiments on a real-world dataset,CelebA.The experimental results illustrate the effectiveness of CCIG.

Tomographic PIV investigation on coherent vortex structures over shark-skin-inspired drag-reducing riblets

Nature has shown us that the microstructure of the skin of fast-swimming sharks in the ocean can reduce the skin friction drag due to the well-known shark-skin effect.In the present study,the effect of shark-skin-inspired riblets on coherent vortex structures in a turbulent boundary layer（TBL） is investigated.This is done by means of tomographic particle image velocimetry（TPIV） measurements in channel fl ws over an acrylic plate of drag-reducing riblets at a friction Reynolds number of 190.The turbulent fl ws over drag-reducing riblets are verifie by a planar time-resolved particle image velocimetry（TRPIV） system initially,and then the TPIV measurements are performed.Two-dimensional（2D） experimental results with a dragreduction rate of around 4.81% are clearly visible over triangle riblets with a peak-to-peak spacing s＋of 14,indicating from the drag-reducing performance that the buffer layer within the TBL has thickened;the logarithmic law region has shifted upward and the Reynolds shear stress decreased.A comparison of the spatial topological distributions of the spanwise vorticity of coherent vortex structures extracted at different wall-normal heights through the improved quadrant splitting method shows that riblets weaken the amplitudesof the spanwise vorticity when ejection（Q2） and sweep（Q4） events occur at the near wall,having the greatest effect on Q4 events in particular.The so-called quadrupole statistical model for coherent structures in the whole TBL is verified Meanwhile,their spatial conditional-averaged topological shapes and the spatial scales of quadrupole coherent vortex structures as a whole in the overlying turbulent fl w over riblets are changed,suggesting that the riblets dampen the momentum and energy exchange between the regions of near-wall and outer portion of the TBL by depressing the bursting events（Q2 and Q4）,thereby reducing the skin friction drag.

Light-Weight Selective Image Encryption for Privacy Preservation

To protect personal privacy and confidential preservation,access control is used to authorize legal users for safe browsing the authorized contents on photos.The access control generates an authorization rule according to each permission assignment.However,the general access control is inappropriate to apply in some social services(e.g.,photos posted on Flickr and Instagram,personal image management in mobile phone) because of the increasing popularity of digital images being stored and managed.With low maintenance loads,this paper integrates the data hiding technique to propose an access control mechanism for privacy preservation.The proposed scheme changes the partial regions of a given image as random pads (called selective image encryption) and only allows the authorized people to remedy the random pads back to meaningful ones which are with similar visual qualities of original ones.

3D image reconstruction with a controllable overlapping number of elemental images in computational integral imaging

In this Letter, we propose a three-dimensional （3D） image reconstruction method with a controllable overlapping number of elemental images in computational integral imaging. The proposed method can control the overlap- ping number of pixels coming from the elemental images by using the subpixel distance based on ray optics between a 3D object and an image sensor. The use of a controllable overlapping number enables us to provide an improved 3D image visualization by controlling the inter-pixel interference within the reconstructed pixels. To find the optimal overlapping number, we simulate the pickup and reconstruction processes and utilize the numerical reconstruction results using a peak signal-to-noise ratio （PSNR） metric. To demonstrate the feasibility of our work in optical experiments, we carry out the preliminary experiments and present the results.

Light controlled imaging probes for intracellular target recognition： photochromism as a tool to enhance sensing precision

Chemosensors and imaging probes have been the focus of significant research interest over the past few decades. In part due to ease of preparation and simplicity in manipulation, fluorescent probes have been extensively used for biomedical applications. When used for #7 vitro cell imaging [1,2],

3D image reconstruction with controllable spatial filtering based on correlation of multiple periodic functions in computational integral imaging

We propose a novel method of slice image reconstruction with controllable spatial filtering by using the correlation of periodic delta-function arrays （PDFAs） with elemental images in computational integral imaging. The multiple PDFAs, whose spatial periods correspond to object＇s depths with the elemental image array （EIA）, can generate a set of spatially filtered EIAs for multiple object depths compared with the conventional method for the depth of a single object. We analyze a controllable spatial filtering effect by the proposed method. To show the feasibility of the proposed method, we carry out preliminary experiments for multiple objects and present the results.

The Image Conversion in Optical Echo-Holography by External Electromagnetic Standing Waves

The recording and playback of information using a reverse stimulated photon—echo hologram when exposed to the recording medium pulse of non-resonant electromagnetic standing wave was considered. It was shown that the spatial intensity distribution in stimulated echo hologram response depended on the electric field intensity of non-resonant standing wave that allowed controlling by a reproducible image.

Measurement, modelling, and closed-loop control of crystal shape distribution： Literature review and future perspectives

Crystal morphology is known to be of great importance to the end-use properties of crystal products, and to affect down-stream processing such as filtration and drying. However, it has been previously regarded as too challenging to achieve automatic closed-loop control. Previous work has focused on controlling the crystal size distribution, where the size of a crystal is often defined as the diameter of a sphere that has the same volume as the crystal. This paper reviews the new advances in morphological population balance models for modelling and simulating the crystal shape distribution （CShD）, measuring and estimating crystal facet growth kinetics, and two- and three-dimensional imaging for on-line characterisation of the crystal morphology and CShD. A framework is presented that integrates the various components to achieve the ultimate objective of model-based closed-loop control of the CShD. The knowledge gaps and challenges that require further research are also identified.

Image lithography in telluride suboxide thin film through controlling “virtual” bandgap

In this work, TeO_(0.7)thin films were prepared by the reactive magnetron-controlling sputtering method. Complex gray-scale patterns were successfully fabricated on TeO_(0.7)thin films through the laser direct writing method.The structural origin of TeO_(0.7)thin film was investigated for gray-scale pattern formation. It is found that multiple gray-scale levels are dependent on the "virtual" bandgap energy of TeO_(0.7)thin films. The bandgap energy changes lead to refractive index and reflectivity difference. Thus, gray-scale tones can be formed. By accurately controlling laser energy, various "virtual" bandgaps can be generated in TeO_(0.7)thin films, and colorful gray-scale levels can be formed. Experimental results indicate that TeO_(0.7)thin film can be used as micro/nano image writing material.

Double regularization control based on level set evolution for tablet packaging image segmentation

This paper proposes a novel double regular- ization control （DRC） method which is used for tablet packaging image segmentation. Since the intensities of tablet packaging images are inhomogenous, it is difficult to make image segmentation. Compared to methods based on level set, the proposed DRC method has some advantages for tablet packaging image segmentation. The local re- gional control term and the rectangle initialization contour are first employed in this method to quickly segment un- even grayscale images and accelerate the curve evolution rate. Gaussian filter operator and the convolution calcula- tion are then adopted to remove the effects of texture noises in image segmentation. The developed penalty energy function, as regularization term, increases the constrained conditions based on the gradient flow conditions. Since the potential function is embedded into the level set of evo- lution equations and the image contour evolutions are bi- laterally extended, the proposed method further improves the accuracy of image contours. Experimental studies show that the DRC method greatly improves the computational efficiency and numerical accuracy, and achieves better results for image contour segmentation compared to other level set methods.