Output Image Research Articles

CRASA: Chili Pepper Disease Diagnosis via Image Reconstruction Using Background Removal and Generative Adversarial Serial Autoencoder

With the recent development of smart farms, researchers are very interested in such fields. In particular, the field of disease diagnosis is the most important factor. Disease diagnosis belongs to the field of anomaly detection and aims to distinguish whether plants or fruits are normal or abnormal. The problem can be solved by binary or multi-classification based on a Convolutional Neural Network (CNN), but it can also be solved by image reconstruction. However, due to the limitation of the performance of image generation, SOTA’s methods propose a score calculation method using a latent vector error. In this paper, we propose a network that focuses on chili peppers and proceeds with background removal through GrabCut. It shows a high performance through an image-based score calculation method. Due to the difficulty of reconstructing the input image, the difference between the input and output images is large. However, the serial autoencoder proposed in this paper uses the difference between the two fake images, instead of the actual input, as a score. We propose a method of generating meaningful images using the GAN structure and classifying three results simultaneously by one discriminator. The proposed method showed a higher performance than previous research, and image-based scores showed the best performance.

Automated urban landscape design: an AI-driven model for emotion-based layout generation and appraisal

The evolution of a city is significantly shaped by the design of its urban landscape. The advancement of artificial intelligence has substantially increased convenience for individuals. This research proposes an urban landscape layout model powered by artificial intelligence that automatically generates urban landscape design based on deep learning (URDDL) with two dimensions: emotional tendency and urban landscape appraisal. The input image represents land use and surrounding road conditions, while the output image depicts the selection of the main entrance and the internal spatial function layout. The Pix2Pix model is trained to learn the internal function layout based on varying land use and road conditions. Additionally, a domain-specific dictionary is constructed using an existing semantic resource vocabulary, where positive and negative sentiment words are compared with their corresponding sentiment values, focusing on categories such as Stimulate, Sense, and Action. Experimental results indicate that the absolute average error of the URDDL model is 91.31%, with a maximum error of 96.87%. The degree of fit is highly appropriate for evaluating the emotional prediction of urban landscapes. The findings demonstrate that the URDDL model outperforms traditional design methods regarding generated results, suggesting its potential for future applications in automated landscape design.

Evaluation of automatic tube current modulation in a CT scanner using a customised homogeneous phantom

Objective.The introduction of automatic tube current modulation (ATCM) has resulted in complex relationships between scanner parameters, patient body habitus, radiation dose, and image quality. ATCM adjusts tube current based on x-ray attenuation variations in the scan region, and overall patient dose depends on a combination of factors. This work aims to develop mathematical models that predict CT radiation dose and image noise in terms of attenuating diameter and all relevant scanner parameters.Approach.A homogenous phantom, equipped with the features to conduct discrete and continuous adaption tests, was developed to model ATCM in a Philips CT scanner. Scanner parameters were varied based on theoretical dose relationships, and a MATLAB script was developed to extract data from DICOM images. R statistical software was employed for data analysis, plotting, and regression modelling.Main Results.Phantom data provided the following insights: Median tube current decreased by 81% as tube potential varied from 80 kVp to 140 kVp. Doubling the DoseRight Index (DRI) from 12 to 24, at 24 cm diameter, produced a 294% increase in mA and a 46% decrease in noise. Mean mA increased by 53% whilst mean noise increased by 5.7% as helical pitch increased from 0.6 to 0.925. Changing rotation time from 0.33s to 0.75s gave a 56% reduction in mean mA and no change in image noise. Increasing detector collimation (n×T) resulted in higher tube currents and lower output image noise values, asnandTwere varied independently. Interpreting these results to apply transformations relevant to each independent variable produced models for tube current and noise with adjusted R-squared values of 0.965 and 0.912, respectively.Significance.The models developed more accurately predict radiation dose and image quality for specific patients and scanner settings. They provide imaging professionals with a practical tool to optimize scan protocols according to patient diameters and clinical objectives.

Robot Grasping Detection Method Based on Keypoints

ABSTRACTThis study introduces a novel keypoint‐based grasp detection network, denoted as GKSCConv‐Net, which operates on n‐channel input images. The network architecture comprises three SCConv2D layers and three SCConvT2D layers. The SCConvT2D layers facilitate upsampling to maintain consistent dimensions between the output and input images. The resultant output consists of maps indicating left grasp points, right grasp points, and grasp center keypoints. The accuracy of predictions is enhanced through the incorporation of the keypoint refinement module and feature fusion module. To validate the model's generalization and applicability, comprehensive training, testing, and evaluation are conducted on diverse data sets, including the Cornell data set, Jacquard data set, and others representing real‐world scenarios. Furthermore, ablation experiments are employed to substantiate the efficacy of the spatial reconstruction unit (SRU) and channel reconstruction unit (CRU) within the SCConv, exploring their impact on grasp keypoint detection outcomes. Real robotic grasping experiments ultimately affirm the model's outstanding performance in practical settings.

Characterization of the Small Robotic Telescope Instrument and Implementation at ITERA Lampung Astronomical Observatory

As time has passed, the technological advances used to run these telescope systems have become faster, more reliable, and more user-friendly. Institut Teknologi Sumatera (ITERA) has installed a small robotic telescope at the Meteorology–Climatology–Geophysics Observing Station in collaboration with an ASTELCO System. The telescope system is designed to make fully robotic observations and can operate in both interactive and unattended robotic modes. This study's primary focus is on the OZT-ALTS telescope robotic system with the optical system using triplet apochromatic lens technology manufactured by LZOS with a focal ratio of f/8 refractor design. The telescope mount is a German Equatorial (NTM-500 manufactured by ASTELCO) with a direct drive system with an absolute encoder, and it allows fully programmable tracking speeds with a typical slewing speed of 20°/sec and a tracking accuracy range of 0.3–1.0 arcsec. For the process of collecting moon observation images, we use Manta G-031B. The OZT-ALTS robotic telescope is now contributing to monitoring crescent moon observation and astronomical studies. This paper presents the scientific motivation behind the OZT-ALTS robotic telescope and the facility's specifications and unique features. Furthermore, we present to characterize the result of the hardware OZT-ALTS robotic telescope system to increase the mechanical and optical performance of the telescope system based on the in-situ test observations. Based on the result, setting the Gain on the camera at 3.5 e-/ADU is the limit for the CCD to produce good-quality output images by utilizing the entire dynamic range without saturation. We also present an example of the science we have performed with the OZT-ALTS robotic telescope. The OZT-ALTS can observe young crescent moon events in the daylight after implementing the CLAHE technique to enhance contrast objects. On the other hand, the OZT-ALTS can also observe WASP-16b and WASP-34b exoplanets by using the transit method to analyze the decreasing light curve when a planet passes between a star and its observer.

A Lightweight Image Encryption based on Improving LED using Rossler Attractor

Lightweight image encryption refers to the use of encryption algorithms designed to secure images while considering the limitations of resources, such as processing power and memory, typically found in devices like IoT devices, mobile phones, and embedded systems. This paper proposes a lightweight image encryption based on an LED encryption algorithm and a method for a pseudo number generator called Rossler attractor. Image is split into their three-color bands (RGB) and then partitioned into a block for encryption, a three-equations model (Rossler Attractor) is used for a pseudo number generator, a specific operation applied for producing key scheduled for rounds in LED to seem as random as possible. The experiment explains the objective test to ensure the performance. The contribution of this work is to present an efficient method for lightweight encryption that is applied to color images with a pseudo number generator-based Rossler attractor dependent on a three-dimensional equation. The technique makes it sensitive to initial conditions and could seem like dynamic encryption that resisted many types of attacks. Visually test represented by the histogram of the image before and after encryption is noted that the result is uniform and differs from the input one. The correlation of the output image is a low value which means no correlation is there between pixels horizontal, vertical and diagonal. The similarity test also applied such as MSE, PSNR and SSIM which also explain no similarity between plain and encrypted images. The randomness of generated numbers was tested using the standard test NIST for evaluation of the input key

Automatic Inbetweening for Stroke‐Based Painterly Animation

AbstractPainterly 2D animation, like the paint‐on‐glass technique, is a tedious task performed by skilled artists, primarily using traditional manual methods. Although CG tools can simplify the creation process, previous works often focus on temporal coherence, which typically results in the loss of the handmade look and feel. In contrast to cartoon animation, where regions are typically filled with smooth gradients, stroke‐based stylized 2D animation requires careful consideration of how shapes are filled, as each stroke may be perceived individually. We propose a method to generate intermediate frames using example keyframes and a motion description. This method allows artists to create only one image for every five to 10 output images in the animation, while the automatically generated intermediate frames provide plausible inbetween frames.

Rectifying inhomogeneous illumination in digital images using a latent light manifestation algorithm

ABSTRACT Rectification of illumination is vital in enhancing the visibility of images acquired in suboptimal lighting conditions. Still, it remains a challenging yet intriguing process as many existing algorithms fail to meet expectations. Hence, a latent light manifestation (LLM) algorithm is introduced, beginning with a conversion to the HSV space, processing the V channel using the Naka-Rushton and exponential equations. Next, the two output images are blended using a two-step range-expansion blending approach. After that, a statistics-based tonality adjustment method and a dynamic range extension process are applied. Finally, a transfer to the RGB space is performed to produce the output image. The LLM is tested using three datasets, compared with eight contemporary algorithms, and the quality evaluations are carried out via five performance measures. The results denote that the developed LLM consistently outperforms existing algorithms, scoring the best according to the used measures and yielding visually pleasing results.

Infrared non-uniformity correction method based on binocular detection system without blindsight

The output image of an infrared detector often suffers from non-uniformity due to the inhomogeneity of the device material and limitations in the manufacturing process, which can reduce the detector's capability to detect and identify targets. It is therefore necessary to adequately correct the non-uniformity to fully utilize the high temperature sensitivity of the detector. Calibration-based non-uniformity correction technology is one of the earliest and most effective methods of correcting non-uniformity. Typically, a relatively uniform shutter is used to mask the field of view for calibration correction. However, this process interrupts the observation due to the need for masking, severely limiting its use in real-world projects. In order to maintain the performance superiority of the calibration-based method and eliminate the defect of target loss caused by masking, this paper takes the occlusion calibration as the theoretical basis, which consists of dual detectors to form a binocular detection system, and uses one of the occluded calibration auxiliary detector to assist the other occlusion-free primary detector, and takes advantage of the fact that the radiation is the same when observing the same target at the same time to make the correct response based on the steepest gradient descent learning algorithm, which removes the non-uniformity while achieving detail fidelity, no ghosting and continuous output.

An intelligent character segmentation system coupled with deep learning based recognition for the digitization of ancient Tamil palm leaf manuscripts

Palm-leaf manuscripts, rich with ancient knowledge in areas such as history, art, and medicine, are vital cultural treasures, making their digitization essential for preserving this heritage. Digitization of these organic and fragile manuscripts is required to safeguard the essential ancient data. This requires optimal character segmentation and recognition algorithms. A limited number of studies have been carried out in Tamil character recognition in literature. Handling row-overlapped characters, noise introduced due to lightning issues, and dirt, as well as the removal of punch holes, auto-cropping the content, filtering out noisy or improper segmentation, etc. are the essential concerns carried out in our proposed work. This work is executed as a four-step process (1) Palm Leaf Manuscript Acquisition (2) Pre-Processing (3) Segmentation of Tamil Characters and (4) Tamil Character Recognition. During acquisition, the scanners are used for recording palm leaf manuscripts from the Tamil Nadu-oriented manuscript library. In the Pre-processing step, the Fast Non-Local Means (Fast-NLM) method, paired with median filtering is used for Denoising the scanner output image. Later, the pixels that make the characters and borders (i.e., the foreground) are identified using Sauvola thresholding. The proposed methodology introduces efficient techniques to remove Punch hole impressions from the pre-processed image, and to crop the written content from the edges. After pre-processing, the Segmentation of Tamil Characters is performed as a three-step process (a) Manuscript (b) Line, and (c) character segmentation, which addresses conjoined lines, partially/completely empty segmentations that are not previously addressed by existing techniques. This work introduces an Augmented HPP line-splitting algorithm that accurately segments written lines, handling wrong segmentation cases that were previously not considered by existing techniques. The system achieves an average segmentation accuracy of 98.25%, which far outperforms existing techniques. It also proposes a novel Punch hole removal algorithm that can locate and remove the punch-hole impressions in the manuscript image. This algorithm, along with the automated content cropping technique, increases recognition accuracy and eliminates any manual labor needed. These features make the proposed methodology highly suitable for real-time archaeological and historical researches that include manuscripts. All 247 letters and 12 numeric digits are analyzed and separated into 125 distinct writable characters. In our work, characters are segmented and used for recognition of all 247 letters and 12 digits in Tamil using a multi-class CNN with 125 classes, which drastically reduces the complexity of the neural network compared to having 257 output nodes. It offered a notable performance of 96.04% accuracy. As compared with existing Tamil and other character recognitions, this work is effective in essence of considering real-time images and the increased number of characters used.

Security of End-to-End medical images encryption system using trained deep learning encryption and decryption network

The Internet of Medical Things (IoMT) links medical devices and wearable, enhancing healthcare. To secure sensitive patient data over the IoMT, encryption is vital to retain confidentiality, prevent tampering, ensure authenticity, and secure data transfer. The intricate neural network architecture of deep learning models adds a layer of complexity and non-linearity to the encryption process, rendering it highly resistant to plaintext attacks. Specifically, the Cycle_GAN network is used as the leading learning network. This work suggests deep learning-based encryption for medical images using Cycle_GAN, a Generative Adversarial Network. Cycle_GAN changes images without paired training data that improves quality and feature preservation. Unlike conventional image-to-image translation techniques, Cycle_GAN doesn’t require a dataset with corresponding input–output pairs. Traditional methods typically needs paired data to learn the mapping between input and output images. Paired data can be challenging to obtain, specifically in medical imaging where gathering and annotating data can be time-consuming, laborious and expensive. The use of Cycle_GAN overwhelms this constraint by using unpaired data, where the input and output images are not explicitly paired. This method ensures confidentiality, authenticity, and secure transfer. Cycle_GAN consists of two major components: a generator used to modify the images, and a discriminator used to distinguish between real and fake images. Further, the Binary-Cross Entropy loss function is employed to train the network for precise predictions. The experiments are carried out on skin cancer datasets. The results demonstrate high-level efficient, systematic and coherent encryption as compared with other modernized medical image encryption methods. The proposed technique offers several benefits, including efficient encryption and decryption and robustness against unauthorized access.

Survival Analysis of Penile Cancer Patients Using Cox-Proportional Hazard and GPT-4

Today, in 2024, the evolution of the evolution of artificial intelligence (AI) is very impressive. An AI product that continues to develop is ChatGPT. The threat of ChatGPT to some job professions is real. Beside that, ChatGPT also helps many professions make their work easier. In this research, we analyze the survival analysis of penile cancer patients. The design of this research is mixed-method (quantitative-qualitative). The quantitative and qualitative methods that were used are Cox-proportional hazards and GPT-4, respectively. Based on quantitative analysis, it can be concluded that the Survival and Hazard model for study penile cancer are and , respectively. The predicted chances of survival and death for the patient in stage 2, 14 days after surgery, are 40% and 60%, respectively. Based on qualitative analysis, GPT-4 cannot quantitatively obtain the cox proportional hazard number. However, even so, GPT4 can read the cox proportional hazard output image, providing Python and R survival analysis coding quite perfectly. Thus, the qualitative method's conclusion is that GPT-4 can greatly assist data analysts in their analysis of survival data.

Attenuated color channel adaptive correction and bilateral weight fusion for underwater image enhancement

Due to the absorption and scattering of light and the influence of suspended particles, underwater images commonly exhibit color distortions, reduced contrast, and diminished details. This paper proposes an attenuated color channel adaptive correction and bilateral weight fusion approach called WLAB to address the aforementioned degradation issues. Specifically, a novel white balance method is first applied to balance the color channel of the input image. Moreover, a local-block-based fast non-local means method is proposed to obtain a denoised version of the color-corrected image. Then, an adaptive stretching method that considers the histogram's local features to get a contrast-enhanced version of the color-corrected image. Finally, a bilateral weight fusion method is proposed to fuse the above two image versions to obtain an output image with complementary advantages. Experimental studies are conducted on three benchmark underwater image datasets and compared with ten state-of-the-art methods. The results show that WLAB has a significant advantage over the comparative methods. Notably, WLAB exhibits a degree of independence from camera settings and enhances the precision of various image processing applications, including key points and saliency detection. Additionally, it demonstrates commendable adaptability in improving low-light and foggy images.

MBHFuse: A multi- branch heterogeneous global and local infrared and visible image fusion with differential convolutional amplification features

Fusing infrared and visible imagery aims to harness their complementary spectral data, enhancing output image quality, sharpness, and content. However, convolutional neural networks (CNNs) are not capable of capturing long-range dependencies, while the Transformer architecture faces the challenge of consuming huge computational resources. To address these critical challenges, this paper proposes a novel framework named MBHFuse, which employs a multi-branch heterogeneous global and local image fusion approach. To effectively extract global and local features, we design a multi-branch heterogeneous encoder module and introduce a differential convolution amplification module (DCAM) to further extract complementary information. Additionally, we devise a new loss function, incorporating multi-branch feature decomposition loss, intensity loss, gradient loss, mean squared error loss, and structural similarity loss, for training the proposed MBHFuse model. Through extensive experiments on public datasets, we demonstrate that the proposed framework outperforms other state-of-the-art (SOTA) methods in both qualitative and quantitative evaluations. Our code will be available at https://github.com/sunyichen1994/MBHFuse.

Objective method of measuring resolution of image intensifier tubes

Image intensifier tubes (IITs) are the most important modules of night vision devices used in huge numbers by military forces worldwide. Resolution is the most important parameter of IITs that presents information about ability of these devices to produce output images preserving information about details of the observed scenery. Despite its importance, it is still a common practice to measure resolution subjectively, by an observer looking at image of a resolution target created by a tested IIT. A series of attempts have been carried out to develop objective methods for accurate resolution measurement of IITs but with limited success. Accuracy of these methods varies depending on the tested IIT. This paper presents detailed analysis of proposed methods for objective resolution measurement. This analysis has shown that significant variability of accuracy of these methods is caused by one main drawback: the methods do not take into account influence of the spatial noise effect on human perception of image of the resolution target. Thus, an improved method taking into account spatial noise and its impact on target detection has been proposed. The method has been validated through experimental verification that shows accuracy improvements compared to other objective methods. This new approach improves accuracy of measurement of resolution of IITs to a level that can be accepted at professional test stations. In this way, this new method has potential to replace the standard subjective method to measure resolution of IITs and fix the biggest flaw of the standard test stations: measurement subjectivity.

Development of deep learning-based novel auto-segmentation for the prostatic urethra on planning CT images for prostate cancer radiotherapy.

Urinary toxicities are one of the serious complications of radiotherapy for prostate cancer, and dose-volume histogram of prostatic urethra has been associated with such toxicities in previous reports. Previous research has focused on estimating the prostatic urethra, which is difficult to delineate in CT images; however, these studies, which are limited in number, mainly focused on cases undergoing brachytherapy uses low-dose-rate sources and do not involve external beam radiation therapy (EBRT). In this study, we aimed to develop a deep learning-based method of determining the position of the prostatic urethra in patients eligible for EBRT. We used contour data from 430 patients with localized prostate cancer. In all cases, a urethral catheter was placed when planning CT to identify the prostatic urethra. We used 2D and 3D U-Net segmentation models. The input images included the bladder and prostate, while the output images focused on the prostatic urethra. The 2D model determined the prostate's position based on results from both coronal and sagittal directions. Evaluation metrics included the average distance between centerlines. The average centerline distances for the 2D and 3D models were 2.07 ± 0.87mm and 2.05 ± 0.92mm, respectively. Increasing the number of cases while maintaining equivalent accuracy as we did in this study suggests the potential for high generalization performance and the feasibility of using deep learning technology for estimating the position of the prostatic urethra.

Rapid detection technology of sewing thread based on DeblurGAN-BSV3 defuzzification algorithm and ST-FPN detection algorithm

Stitch defect detection is an important means to ensure the quality of stitched fabrics. In order to realize the intellectualization of the sewing process, a rapid defect detection method based on machine vision was proposed. In order to solve the problem of motion blur caused by sewing speed, an image DeblurGAN-BSV3 algorithm is proposed. Firstly, different attention fusion mechanisms are introduced to achieve accurate extraction of image features. Secondly, the depth-separable convolution of the trunk network is replaced by blueprint convolution to ensure the processing speed of the detection algorithm and the clarity of the output image. In order to solve the difficulty of rapid detection of pin defects, a student–teacher feature pyramid matching method was proposed. Firstly, the abnormal region interpretation technology was used to explicitly visualize the defect area of the identified abnormal instances, and then the abnormal detection score was performed on the detected images to make a quality judgment. The experimental results show that the combination of the above two methods has a good ability for rapid defect detection. Rapid detection of common pin defects is realized.

Robust phase extraction from complex signals/fringe patterns for optical communications in 5G and beyond through continuous wavelet ridge technique

AbstractThe rapid progress in high‐precision optical instruments necessitates sophisticated image processing techniques for extracting vital information from generated images. Most of these instruments output images comprising RGB pixels. Deciphering these pixel values into phase information poses a significant challenge, especially in the presence of background noise. This study focuses on utilizing Two‐Dimensional Continuous Wavelet Transforms (2‐D CWT) for analyzing fringe patterns with varying noise levels and fringe alignments, crucial for high‐precision optical systems pivotal in enhancing the performance and reliability of optical communication systems in advanced 5G networks. The simulation results demonstrate that 2‐D CWT efficiently extracts phase information from complex and highly noisy fringes while requiring less computational time. Furthermore, the algorithm effectively handles noise disturbances with a commendable degree of accuracy, ensuring robust performance in 5G‐enabled optical systems critical for supporting ultra‐high‐speed data transmission and low‐latency communication requirements. This research contributes to optimizing image analysis techniques for 5G optical systems, facilitating their integration into next‐generation communication networks.

Curriculum learning for ab initio deep learned refractive optics

Deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. However, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a DeepLens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. We demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.

Medical X-ray images enhancement based on super resolution convolution neural network

<p>Pneumonia is a severe lung infection, chest X-ray (CXR) image preferred to find infection. Real images lost its quality, resolution and other feature due to transmission. So good qualitative datasets are very limited. Quality enhancement in medical images is challenging task for researchers. And quality in clinical diagnosis of any disease in deep learning play a very important role. So, this paper presents an aspect with importance of quality in medical images CXR of a particular dataset and how to enhance and create new images with high quality resolution, that is re-used for classification in deep learning. Super resolution convolutional neural netwok (SRCNN) is deep learning based method, which is used for improving resolution in image. Super resolution means low resolution (LR) images from dataset is to be reconstructed or magnified into high resolution (HR). The objective behind this study is to measure the effect of super resolution with quality index, peak signal-to-noise ratio (PSNR), mean squared error (MSE), and structural similarity index measure (SSIM). This experinment performed on 200 images with 10 batches, each batch has 20 images from Kermany dataset, select LR images and converted into HR with SRCNN. Then we find PSNR value of image is increase upto 2 to 5 DB, and MSE of qood quality images is near to zero and MSE decrease up to 20-25, SSIM value have little variation due to same pattern is found in input and output images. Enhancement means highlight or improve the region of interest of pneumonic images. Main goal of this study is to preapare a modified dataset which is further used for classification.</p>