Image Transformation Research Articles

Image Transformation Through Pencil Art Rendering and Selective Colorization

Image Transformation Through Pencil Art Rendering and Selective Colorization

Detecting severity of Diabetic Retinopathy from fundus images: A transformer network-based review

Diabetic Retinopathy (DR) is considered one of the significant concerns worldwide, primarily due to its impact on causing vision loss among most people with diabetes. The severity of DR is typically comprehended manually by ophthalmologists from fundus photography-based retina images. This paper deals with an automated understanding of the severity stages of DR. In the literature, researchers have focused on this automation using traditional machine learning-based algorithms and convolutional architectures. However, the past works hardly focused on essential parts of the retinal image to improve the model performance. In this study, we adopt and fine-tune transformer-based learning models to capture the crucial features of retinal images for a more nuanced understanding of DR severity. Additionally, we explore the effectiveness of image transformers to infer the degree of DR severity from fundus photographs. For experiments, we utilized the publicly available APTOS-2019 blindness detection dataset, where the performances of the transformer-based models were quite encouraging.

Detection of Shoreline Changes With Automated Water Extraction Index (AWEI) In Galur District, Kulon Progo Regency

Abstract Monitoring shoreline change is one of the efforts in the management and protection of the coastal environment. Through digital image processing as the utilization of shoreline extraction index can be used as a reference in measuring the rate of shoreline change. Galur District is one of the sub-districts in Kulon Progo Regency that experiences the highest abrasion process, the worst abrasion that has occurred in recent years is located at Trisik Beach where abrasion threatens the existence of stalls, turtle conservation sites and also coastal vegetation that grows along the beach. The purpose of this study is to determine changes in the coastline starting in 2013, 2018, and 2023 and also to analyze numerically through DSAS the rate of change in the coastline both the influence of abrasion and accretion processes in Galur District. The method used in this research is Quantitative Descriptive assisted by Landsat 8 image transformation using Automated Water Extraction Index (AWEI) and measuring the rate of shoreline change using DSAS based on EPR and NSM values. The results of this study show that Banaran Village has the highest shoreline retreat changes where in 2013 - 2018 the shoreline retreat reached 70 meters and in 2018 - 2023 the coastal retreat reached -75.7 meters, while for the rate of change in Karang Sewu Village in 2013 - 2018 the shoreline retreat reached -33.2 meters and in 2018 - 2023 showed the highest abrasion of -33.5 meters.

Multi-channel and multi-featured extended orb with gabor filter and non-maximum suppression

This research paper presents an enhanced approach to feature extraction using the Oriented FAST and Rotated BRIEF (ORB) algorithm (Rublee et al., 2014). The proposed method leverages the power of Gabor filters (Mehrotra et al., 1992) in combination with Non-Maximum Suppression (Neuback and Van Gool, 2006) to improve the robustness and efficiency of feature detection in computer vision applications. The process begins with the acquisition of an RGB image, which is then transformed into seven single-channel images representing different color and intensity aspects: red, green, blue, hue, saturation, value, and grayscale. Each single-channel image is independently subjected to Gabor filtering, resulting in seven Gabor-filtered images. These images capture distinct texture and frequency information, enhancing the discriminative power of subsequent feature extraction. Subsequently, the Oriented FAST and Rotated BRIEF (ORB) algorithm is applied to each of the seven Gabor-filtered images to extract key points and descriptors. This step yields a comprehensive set of keypoints (Mallick et al., 2015) and descriptors, capturing rich local feature information across multiple image channels. To further refine the extracted features and eliminate redundant keypoints, Non-Maximum Suppression is employed independently on each single-channel image. This process effectively filters out overlapped and false keypoints, ensuring the selection of only the most salient features for downstream tasks. Experimental results demonstrate the efficacy of the proposed approach in improving feature detection performance compared to traditional methods. The combination of Gabor filtering and Non-Maximum Suppression enhances feature discriminability, robustness to noise, and resistance to image transformations, making it well-suited for various computer vision applications, including object recognition, image matching, and scene understanding.

GENRE-STYLE CONSTANTS OF CHAMBER SYMPHONY BY E. STANKOVYCH

The originality of the creative handwriting of Ukrainian composers of the end of the ХХ and the first quarter of the ХХІ century is becoming an urgent research problem of our time. A large number of components of the musical language of artistic compositions, understanding of new value orientations, genre-style synthesis - these and other features constitute the poetics of the work of domestic artists of the mentioned period. E. Stankovych took an active place among the constellation of Ukrainian composers of the end of the 20th and the first quarter of the 21st century. The article emphasizes that the chamber symphony genre in the composer's work shows the peculiarities of his creative handwriting. It was determined that the composer's legacy of E. Stankovych, in particular the study of his chamber symphonies, is a production for the training of teachers of artistic disciplines. After all, chamber symphonies reflect constant reforms in the field of musical language and composer's thinking. The purpose of the research is to identify genre and stylistic features of chamber symphonies by E. Stankovych. To achieve the goal, historiographical, genre-stylistic and comparative research methods were used. The genre-style model of E. Stankovich's chamber symphonies shows the specificity of postmodern art, which takes place in genre synthesis, understanding the semantic potential of timbre, and transformation of images. Fifteen chamber symphonies of E. Stankovych have a common principle of construction, which gives grounds for confirming the opinion about the emergence of the author's metastyle.

Crater Triangle Matching Algorithm Based on Fused Geometric and Regional Features

Craters are regarded as significant navigation landmarks during the descent and landing process in small body exploration missions for their universality. Recognizing and matching craters is a crucial prerequisite for visual and LIDAR-based navigation tasks. Compared to traditional algorithms, deep learning-based crater detection algorithms can achieve a higher recognition rate. However, matching crater detection results under various image transformations still poses challenges. To address the problem, a composite feature-matching algorithm that combines geometric descriptors and region descriptors (extracting normalized region pixel gradient features as feature vectors) is proposed. First, the geometric configuration map is constructed based on the crater detection results. Then, geometric descriptors and region descriptors are established within each feature primitive of the map. Subsequently, taking the salience of geometric features into consideration, composite feature descriptors with scale, rotation, and illumination invariance are generated through fusion geometric and region descriptors. Finally, descriptor matching is accomplished by computing the relative distances between descriptors and adhering to the nearest neighbor principle. Experimental results show that the composite feature descriptor proposed in this paper has better matching performance than only using shape descriptors or region descriptors, and can achieve a more than 90% correct matching rate, which can provide technical support for the small body visual navigation task.

Strobe training as a visual training method that improves performance in climbing.

Strobe training is a form of visual training where the athlete has to practice during intermittently dark conditions. Strobe training improves visual, perceptual, and cognitive skills, which will enhance athletic performance. Strobe training can influence multiple training components in climbing: psychological, tactical, physical, and technical training. The study was conducted on 17 elite climbers from Romania (10 male and 7 female), representing the entire National Youth Climbing Team. The research group was divided into a control group (n = 8) and an experimental group (n = 9). The used instruments were the Cognitrom battery (for cognitive skills, such as spatial skills and reactivity), the Witty SEM system (for motor-cognitive skills, such as cognitive agility, visual processing speed, and visual memory), and the International Rock Climbing Research Association (IRCRA) performance-related test battery for climbers (climbing-specific motor skills). The experimental group had 20 strobe training sessions, which took place during one calendar year, as an additional session to their climbing schedule done with their principal trainer. The strobe session was once a week, depending on the periodization of the macrocycle (preparatory, competitional, and transition periods). The control group and the experimental group had similar climbing training sessions during the 1-year macrocycle in terms of intensity and volume of their training. Strobe training improved on-sight performance (d = 0.38) and red-point performance (d = 0.36). Strobe training improved the majority of cognitive skills [all spatial skills (d = 1.27 for mental image transformation; d = 1.14 for spatial orientation; d = 1.59 for image generation) and simple reaction time (d = 0.99)]. Strobe training improved all motor-cognitive skills (d = 0.16 for visual memory; d = 1.96 for visual memory errors; d = 1.39 for visual processing speed; d = 1.94 for visual processing errors; d = 1.30 for cognitive agility). Strobe training improved many climbing-specific parameters (flexibility and upper body strength) (d = 0.44 and d = 0.47 for flexibility parameters; d = 0.50 to 0.73 for upper body strength parameters). Strobe training is an effective training method for enhancing performance that should be used on more experienced climbers. It acts more on spatial skills, rather than on reactivity skills, developing the visual-motor coordination system. Strobe training has greater effects on climbers aged below 16 years, as youth athletes rely more on visual input compared to adults. The improvement in climbing-specific variables was due to the additional climbing session done weekly. Strobe training acts more on the cognitive component of training than on the motor component of training in climbing.

PERCEPTION OF RUSSIAN PANSLAVISM IN THE BRITISH PARLIAMENTARY DEBATES N THE BEGINNING OF THE EASTERN CRISIS OF THE 1870S

During the Eastern Crisis of the 1870s, the British political elites followed with particular attention the Panslavic ideas expressed in Russia and often perceived them as a threat of the Russian Empire’s expansion into the Balkan Peninsula. However, at the very beginning of the Eastern Crisis, the British Parliament’s attitude to the Russian Panslavism was neutral or even positive and it had a considerable impact on British-Russian relations. Britain’s perception of Russian Panslavism during this short but significant period has not been researched yet. This article uses the materials of parliamentary debates to trace the transformation of images of Russian Panslavism in the British political discourse in 1876 and to analyze the mechanisms of their construction. The author identifies three models of interpretation of Russia’s “Slavic interests”, first, as “Slavic sympathies” of the Russian population, second, as Russia’s “territorial ambitions”, and third, as “Russian intrigues”, which allegedly instigated Slav uprisings against the Ottoman Empire. In the summer of 1876, the interpretation of Panslavism as sincere sympathies of Russian society for its Slavic brothers received the greatest weight in the British political discourse. This interpretation contributed to the formation of a more positive image of Russia and even a temporary change in the pro-Turkish policy of the British government. However, by the late 1876, British politicians increasingly began to suspect the Russian government of “subversive Panslavism”. The formation of images of Russian Panslavism in the political discourse of Great Britain was influenced not only the traditional patterns of British foreign policy, but also by the internal political struggle that escalated during the “Bulgarian agitation”. Panslavism was also interpreted ambiguously due to the lack of information, so politicians often based their judgments on stereotypes about Russia, the Ottoman Empire and the Balkan peoples. Finally, parliamentarians’ views were also affected by a rethinking of their own British identity. British politicians revisited issues of relevance to Britain’s own attitudes to the sovereignty of empires, the uprisings of colonized peoples, moral responsibility and Britain’s international mission.

The picture is in the collection “The Suggestion of the Wounded and the Injury of the Wounded” by Al-Husri Al-Qayrawani

The picture is in the collection “The Suggestion of the Wounded and the Injury of the Wounded” by Al-Husri Al-Qayrawani

Fabrication of Gradient Periodic Surface Structures on Stainless Steel Using Direct Laser Interference Patterning

In this work, the fabrication of gradient periodic line‐like structures with continuously changing spatial period on stainless steel by means of two‐beam direct laser interference patterning (DLIP) is studied. A 70 ps (picosecond [ps]) pulsed laser operating at 532 nm wavelength and 0.64 J cm−2 laser fluence is used to produce the gradient microstructures. By continuously changing the overlapping angle between the sub‐beams, the spatial period of gradient structures smoothly increases from 2.01 ± 0.09 to 3.99 ± 0.12 μm within a 500 μm long processed area. Different pulse‐to‐pulse overlaps are applied to study the influence of cumulated fluence on the topography of gradient structures. The surface topography is analyzed by optical confocal microscopy and is characterized by the roughness parameters Ra and Rz. Fourier transformations of the confocal images allows the verification of the variations of the spatial periods. In the experiments performed in this work, design rules are provided for fabricating complex microstructures in the future using DLIP‐based technology to achieve advanced surface function like anisotropic wetting or cell growth.

Self-supervised monocular visual odometry based on cross-correlation

Visual odometry constitutes a critical component in enabling autonomous navigation. However, the existing methods are limited by the feature extraction and matching accuracy, and cannot show good real-time performance while combining accuracy and robustness. In this paper, we propose a novel monocular visual odometry framework based on cross-correlation. The framework starts with a parameter-sharing Siamese network to build feature extractors that can simultaneously process multiple images as inputs. Moreover, we design cross-correlation modules and define a cross-correlation matrix to describe the strength of correlation between different parts of the input feature maps, reflecting the rotational and translational transformations of the input images. Furthermore, a novel loss function is introduced to impose constraints on the network. Additionally, a fully convolutional network is designed for pose estimation, computing poses alterations from the structure of the cross-correlation matrix. Channel attention and spatial attention mechanisms are introduced to improve the performance. More importantly, our method innovatively uses time intervals as labels, enables self-supervised training, and relies only on a monocular camera. Experimental results on the KITTI visual odometry dataset and the Oxford Robotcar Dataset show that our method produces competitive performance, demonstrating the superiority of the proposed method.

Evaluation of the Degree of Degradation of Brake Pad Friction Surfaces Using Image Processing

The improvement of drilling rig systems to ensure a reduction in unproductive time spent on lowering and lifting operations for replacing drilling tools and restoring the performance of drilling equipment units is an important task. At the same time, considerable attention is paid to the reliable and efficient operation of the braking systems of drilling rig winches. In the process of operation, the polymer pads periodically come into contact with the outer cylindrical surface of the metal pulley during braking, work in extreme conditions and wear out intensively, so they need periodic replacement. Tests were carried out on a modernized stand and in industrial conditions for the brakes of drilling winches. A methodology for evaluating the degradation of the brake pad friction surface during its operation is proposed. The assessment of the degradation degree is carried out based on the image of the brake pad surface using image processing techniques. Geometric transformations of the input image were performed to avoid perspective distortions caused by the concave shape of the brake pads and the spatial angle at which the image is acquired to avoid glares. The crack detection step was implemented based on the scale-space theory, followed by contour detection and skeletonization. The ratios of the area and perimeter of segmented and skeletonized cracks to the total area were chosen as integral characteristics of the degradation degree. With the help of scanning electron microscopy, the character of the destruction of the friction surface and the degradation of the polymer material was investigated. Experimental studies were performed, and the application of the proposed method is illustrated.

Women, Film and Politics in the Islamic Republic of Iran 1979-2005; A Perspective from Learning from Experience and Performance Theory

Images in popular culture provide essential information about the socio-political situation and the self-realisation of gender roles in society. The transformation of women characters in Iranian films after the 1980s has occupied a certain space in scholarship, however everyday self-experiences in the streets and its relation to self realization and women figures are barely included into the analysis. Based on learning from experience, affect and performance theory, this paper focesses on two prominent female stars of Iranian cinema: Niki Karimi and Hediye Tehrani as image transformers in a period of cultural evolution and the smile of women that fades away from the streets. The analysis begins with a general examination of the interaction between politics and culture through the main currents in Iranian cinema. In this part, a periodization of the post-revolution era in Iran is presented through cultural products, especially Iranian films such as the religious era, the war and post-war era, the pastoral era, and the critical era. The last part compares two famous women's images, Niki Karimi and Hediye Tehrani, and shows how this periodization helps to understand the changes in women’s self realization and body politics in everyday life after the revolution.

Segmenting medical images with limited data

While computer vision has proven valuable for medical image segmentation, its application faces challenges such as limited dataset sizes and the complexity of effectively leveraging unlabeled images. To address these challenges, we present a novel semi-supervised, consistency-based approach termed the data-efficient medical segmenter (DEMS). The DEMS features an encoder–decoder architecture and incorporates the developed online automatic augmenter (OAA) and residual robustness enhancement (RRE) blocks. The OAA augments input data with various image transformations, thereby diversifying the dataset to improve the generalization ability. The RRE enriches feature diversity and introduces perturbations to create varied inputs for different decoders, thereby providing enhanced variability. Moreover, we introduce a sensitive loss to further enhance consistency across different decoders and stabilize the training process. Extensive experimental results on both our own and three public datasets affirm the effectiveness of DEMS. Under extreme data shortage scenarios, our DEMS achieves 16.85% and 10.37% improvement in dice score compared with the U-Net and top-performed state-of-the-art method, respectively. Given its superior data efficiency, DEMS could present significant advancements in medical segmentation under small data regimes. The project homepage can be accessed at https://github.com/NUS-Tim/DEMS.

Combining EfficientNet with ML-Decoder classification head for multi-label retinal disease classification

Retinal diseases that are not treated in time can cause irreversible, permanent damage, including blindness. Although a patient may suffer from more than one retinal disease at the same time, most of the studies focus on the diagnosis of a single disease only. Therefore, to detect multi-label retinal diseases from color fundus images, we developed an end-to-end deep learning architecture that combines the EfficientNet backbone with the ML-Decoder classification head in this study. While EfficientNet provides powerful feature extraction with fewer parameters via compound scaling, ML-Decoder further improves efficiency and flexibility by reducing quadratic dependency to a linear one and using a group decoding scheme. Also, with the use of sharpness-aware minimization (SAM) optimizer, which minimizes loss value and loss sharpness simultaneously, higher accuracy rates have been reached. In addition, a significant increase in EfficientNet performance is achieved by using image transformations and concatenation together. During the training phase, the random application of the image transformations allows for increasing the image diversity and makes the model more robust. Besides, fusing fundus images of left and right eyes at the pixel level extracts useful information about their relationship. The performance of the final model was evaluated on the publicly available Ocular Disease Intelligent Recognition (ODIR) dataset consisting of 10,000 fundus images, and superior results were obtained in all test set scenarios and performance metrics than state-of-the-art methods. The best results we obtained in the threefold cross-validation scenario for the kappa, F1, and AUC scores are 68.96%, 92.48%, and 94.80%, respectively. Moreover, it can be considered attractive in terms of floating point operations per second (FLOP) and a number of parameters.

Unsupervised disentanglement strategy for mitigating artifact in photoacoustic tomography under extremely sparse view

Traditional methods under sparse view for reconstruction of photoacoustic tomography (PAT) often result in significant artifacts. Here, a novel image to image transformation method based on unsupervised learning artifact disentanglement network (ADN), named PAT-ADN, was proposed to address the issue. This network is equipped with specialized encoders and decoders that are responsible for encoding and decoding the artifacts and content components of unpaired images, respectively. The performance of the proposed PAT-ADN was evaluated using circular phantom data and the animal in vivo experimental data. The results demonstrate that PAT-ADN exhibits excellent performance in effectively removing artifacts. In particular, under extremely sparse view (e.g., 16 projections), structural similarity index and peak signal-to-noise ratio are improved by ∼188 % and ∼85 % in in vivo experimental data using the proposed method compared to traditional reconstruction methods. PAT-ADN improves the imaging performance of PAT, opening up possibilities for its application in multiple domains.

The anterior cingulate cortex is involved in intero-exteroceptive integration for spatial image transformation of the self-body

Spatial image transformation of the self-body is a fundamental function of visual perspective-taking. Recent research underscores the significance of intero-exteroceptive information integration to construct representations of our embodied self. This raises the intriguing hypothesis that interoceptive processing might be involved in the spatial image transformation of the self-body. To test this hypothesis, the present study used functional magnetic resonance imaging to measure brain activity during an arm laterality judgment (ALJ) task. In this task, participants were tasked with discerning whether the outstretched arm of a human figure, viewed from the front or back, was the right or left hand. The reaction times for the ALJ task proved longer when the stimulus presented orientations of 0°, 90°, and 270° relative to the upright orientation, and when the front view was presented rather than the back view. Reflecting the increased reaction time, increased brain activity was manifested in a cluster centered on the dorsal anterior cingulate cortex (ACC), suggesting that the activation reflects the involvement of an embodied simulation in ALJ. Furthermore, this cluster of brain activity exhibited overlap with regions where the difference in activation between the front and back views positively correlated with the participants' interoceptive sensitivity, as assessed through the heartbeat discrimination task, within the pregenual ACC. These results suggest that the ACC plays an important role in integrating intero-exteroceptive cues to spatially transform the image of our self-body.

Image Domain Multi-Material Decomposition Noise Suppression Through Basis Transformation and Selective Filtering.

Spectral CT can provide material characterization ability to offer more precise material information for diagnosis purposes. However, the material decomposition process generally leads to amplification of noise which significantly limits the utility of the material basis images. To mitigate such problem, an image domain noise suppression method was proposed in this work. The method performs basis transformation of the material basis images based on a singular value decomposition. The noise variances of the original spectral CT images were incorporated in the matrix to be decomposed to ensure that the transformed basis images are statistically uncorrelated. Due to the difference in noise amplitudes in the transformed basis images, a selective filtering method was proposed with the low-noise transformed basis image as guidance. The method was evaluated using both numerical simulation and real clinical dual-energy CT data. Results demonstrated that compared with existing methods, the proposed method performs better in preserving the spatial resolution and the soft tissue contrast while suppressing the image noise. The proposed method is also computationally efficient and can realize real-time noise suppression for clinical spectral CT images.

Joint Coverless Steganography and Image Transformation for Covert Communication of Secret Messages

Joint Coverless Steganography and Image Transformation for Covert Communication of Secret Messages

SSiT: Saliency-Guided Self-Supervised Image Transformer for Diabetic Retinopathy Grading.

Self-supervised Learning (SSL) has been widely applied to learn image representations through exploiting unlabeled images. However, it has not been fully explored in the medical image analysis field. In this work, Saliency-guided Self-Supervised image Transformer (SSiT) is proposed for Diabetic Retinopathy (DR) grading from fundus images. We novelly introduce saliency maps into SSL, with a goal of guiding self-supervised pre-training with domain-specific prior knowledge. Specifically, two saliency-guided learning tasks are employed in SSiT: 1) Saliency-guided contrastive learning is conducted based on the momentum contrast, wherein fundus images' saliency maps are utilized to remove trivial patches from the input sequences of the momentum-updated key encoder. Thus, the key encoder is constrained to provide target representations focusing on salient regions, guiding the query encoder to capture salient features. 2) The query encoder is trained to predict the saliency segmentation, encouraging the preservation of fine-grained information in the learned representations. To assess our proposed method, four publicly-accessible fundus image datasets are adopted. One dataset is employed for pre-training, while the three others are used to evaluate the pre-trained models' performance on downstream DR grading. The proposed SSiT significantly outperforms other representative state-of-the-art SSL methods on all downstream datasets and under various evaluation settings. For example, SSiT achieves a Kappa score of 81.88% on the DDR dataset under fine-tuning evaluation, outperforming all other ViT-based SSL methods by at least 9.48%.