Journal of Imaging最新文献

{"title":"Enhancing Breast Lesion Detection in Mammograms via Transfer Learning.","authors":"Beibit Abdikenov, Dimash Rakishev, Yerzhan Orazayev, Tomiris Zhaksylyk","doi":"10.3390/jimaging11090314","DOIUrl":"10.3390/jimaging11090314","url":null,"abstract":"<p><p>Early detection of breast cancer via mammography enhances patient survival rates, prompting this study to assess object detection models-Cascade R-CNN, YOLOv12 (S, L, and X variants), RTMDet-X, and RT-DETR-X-for detecting masses and calcifications across four public datasets (INbreast, CBIS-DDSM, VinDr-Mammo, and EMBED). The evaluation employs a standardized preprocessing approach (CLAHE, cropping) and augmentation (rotations, scaling), with transfer learning tested by training on combined datasets (e.g., INbreast + CBIS-DDSM) and validating on held-out sets (e.g., VinDr-Mammo). Performance is measured using precision, recall, mean Average Precision at IoU 0.5 (mAP50), and F1-score. YOLOv12-L excels in mass detection with an mAP50 of 0.963 and F1-score up to 0.917 on INbreast, while RTMDet-X achieves an mAP50 of 0.697 on combined datasets with transfer learning. Preprocessing improves mAP50 by up to 0.209, and transfer learning elevates INbreast performance to an mAP50 of 0.995, though it incurs 5-11% drops on CBIS-DDSM (0.566 to 0.447) and VinDr-Mammo (0.59 to 0.5) due to domain shifts. EMBED yields a low mAP50 of 0.306 due to label inconsistencies, and calcification detection remains weak (mAP50 < 0.116), highlighting the value of high-capacity models, preprocessing, and augmentation for mass detection while identifying calcification detection and domain adaptation as key areas for future investigation.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470960/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maternal Factors, Breast Anatomy, and Milk Production During Established Lactation-An Ultrasound Investigation.","authors":"Zoya Gridneva, Alethea Rea, David Weight, Jacki L McEachran, Ching Tat Lai, Sharon L Perrella, Donna T Geddes","doi":"10.3390/jimaging11090313","DOIUrl":"10.3390/jimaging11090313","url":null,"abstract":"<p><p>Obesity is linked to suboptimal breastfeeding outcomes, yet the relationships between maternal adiposity, breast anatomy, and milk production (MP) have not been investigated. We conducted ultrasound imaging to assess the breast anatomy of 34 lactating women. The amount of glandular tissue (glandular tissue representation (GTR)) was classified as low, moderate, or high. Number and diameters of main milk ducts and mammary blood flow (resistive index) were measured. Women completed a 24 h MP measurement and an obstetric/lactation history questionnaire. Body composition was measured with bioimpedance spectroscopy. Statistical analysis employed correlation networks. Multiple relationships were revealed, with later menarche correlating with minimal pubertal and pregnancy breast growth. A minimal breast growth was further correlated with lower mammary blood flow during lactation and lower numbers and smaller diameters of main milk ducts, which in turn correlated with a lower MP. Importantly, higher adiposity also correlated with minimal breast growth during pregnancy and low GTR and MP. Several modifiable and non-modifiable maternal factors may be associated with breast development and MP. Antenatal lactation assessment and intervention in high-risk women may ensure they reach their full lactation potential and inform future interventions, such as maintaining healthy adiposity.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470485/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EFIMD-Net: Enhanced Feature Interaction and Multi-Domain Fusion Deep Forgery Detection Network.","authors":"Hao Cheng, Weiye Pang, Kun Li, Yongzhuang Wei, Yuhang Song, Ji Chen","doi":"10.3390/jimaging11090312","DOIUrl":"10.3390/jimaging11090312","url":null,"abstract":"<p><p>Currently, deepfake detection has garnered widespread attention as a key defense mechanism against the misuse of deepfake technology. However, existing deepfake detection networks still face challenges such as insufficient robustness, limited generalization capabilities, and a single feature extraction domain (e.g., using only spatial domain features) when confronted with evolving algorithms or diverse datasets, which severely limits their application capabilities. To address these issues, this study proposes a deepfake detection network named EFIMD-Net, which enhances performance by strengthening feature interaction and integrating spatial and frequency domain features. The proposed network integrates a Cross-feature Interaction Enhancement module (CFIE) based on cosine similarity, which achieves adaptive interaction between spatial domain features (RGB stream) and frequency domain features (SRM, Spatial Rich Model stream) through a channel attention mechanism, effectively fusing macro-semantic information with high-frequency artifact information. Additionally, an Enhanced Multi-scale Feature Fusion (EMFF) module is proposed, which effectively integrates multi-scale feature information from various layers of the network through adaptive feature enhancement and reorganization techniques. Experimental results show that compared to the baseline network Xception, EFIMD-Net achieves comparable or even better Area Under the Curve (AUC) on multiple datasets. Ablation experiments also validate the effectiveness of the proposed modules. Furthermore, compared to the baseline traditional two-stream network Locate and Verify, EFIMD-Net significantly improves forgery detection performance, with a 9-percentage-point increase in Area Under the Curve on the CelebDF-v1 dataset and a 7-percentage-point increase on the CelebDF-v2 dataset. These results fully demonstrate the effectiveness and generalization of EFIMD-Net in forgery detection. Potential limitations regarding real-time processing efficiency are acknowledged.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12471254/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151284","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cascaded Spatial and Depth Attention UNet for Hippocampus Segmentation.","authors":"Zi-Zheng Wei, Bich-Thuy Vu, Maisam Abbas, Ran-Zan Wang","doi":"10.3390/jimaging11090311","DOIUrl":"10.3390/jimaging11090311","url":null,"abstract":"<p><p>This study introduces a novel enhancement to the UNet architecture, termed Cascaded Spatial and Depth Attention U-Net (CSDA-UNet), tailored specifically for precise hippocampus segmentation in T1-weighted brain MRI scans. The proposed architecture integrates two key attention mechanisms: a Spatial Attention (SA) module, which refines spatial feature representations by producing attention maps from the deepest convolutional layer and modulating the matching object features; and an Inter-Slice Attention (ISA) module, which enhances volumetric uniformity by integrating related information from adjacent slices, thereby reinforcing the model's capacity to capture inter-slice dependencies. The CSDA-UNet is assessed using hippocampal segmentation data derived from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Decathlon, two benchmark studies widely employed in neuroimaging research. The proposed model outperforms state-of-the-art methods, achieving a Dice coefficient of 0.9512 and an IoU score of 0.9345 on ADNI and Dice scores of 0.9907/0.8963 (train/validation) and an IoU score of 0.9816/0.8132 (train/validation) on the Decathlon dataset across multiple quantitative metrics. These improvements underscore the efficacy of the proposed dual-attention framework in accurately explaining small, asymmetrical structures such as the hippocampus, while maintaining computational efficiency suitable for clinical deployment.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470259/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of AI Performance in Spinal Radiographic Measurements Compared to Radiologists: A Study of Accuracy and Efficiency.","authors":"Francesco Pucciarelli, Guido Gentiloni Silveri, Marta Zerunian, Domenico De Santis, Michela Polici, Antonella Del Gaudio, Benedetta Masci, Tiziano Polidori, Giuseppe Tremamunno, Raffaello Persechino, Giuseppe Argento, Marco Francone, Andrea Laghi, Damiano Caruso","doi":"10.3390/jimaging11090310","DOIUrl":"10.3390/jimaging11090310","url":null,"abstract":"<p><p>This study aimed to evaluate the reliability of an AI-based software tool in measuring spinal parameters-Cobb angle, thoracic kyphosis, lumbar lordosis, and pelvic obliquity-compared to manual measurements by radiologists and to assess potential time savings. In this retrospective monocentric study, 56 patients who underwent full-spine weight-bearing X-rays were analyzed. Measurements were independently performed by an experienced radiologist, a radiology resident, and the AI software. A consensus between two senior experts established the ground truth. Lin's Concordance Correlation Coefficient (CCC), mean absolute error (MAE), ICC, and paired <i>t</i>-tests were used for statistical analysis. The AI software showed excellent agreement with human readers (CCC > 0.9) and demonstrated lower MAE than the resident in Cobb angle and lumbar lordosis measurements but slightly underperformed in thoracic kyphosis and pelvic obliquity. Importantly, the AI significantly reduced analysis time compared to both the experienced radiologist and the resident (<i>p</i> < 0.001). These findings suggest that the AI tool offers a reliable and time-efficient alternative to manual spinal measurements and may enhance accuracy for less experienced radiologists.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470944/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Hybrid Framework for Red Blood Cell Labeling Using Elliptical Fitting, Autoencoding, and Data Augmentation.","authors":"Bundasak Angmanee, Surasak Wanram, Amorn Thedsakhulwong","doi":"10.3390/jimaging11090309","DOIUrl":"10.3390/jimaging11090309","url":null,"abstract":"<p><p>This study aimed to develop a local dataset of abnormal RBC morphology from confirmed cases of anemia and thalassemia in Thailand, providing a foundation for medical image analysis and future AI-assisted diagnostics. Blood smear samples from six hematological disorders were collected between April and May 2025, with twelve regions of interest segmented into approximately 34,000 single-cell images. To characterize cell variability, a convolutional autoencoder was applied to extract latent features, while ellipse fitting was used to quantify cell geometry. Expert hematologists validated representative clusters to ensure clinical accuracy, and data augmentation was employed to address class imbalance and expand rare morphological types. From the dataset, 14,089 high-quality single-cell images were used to classify RBC morphology into 36 clinically meaningful categories. Unlike existing datasets that rely on limited or curated samples, this dataset reflects population-specific characteristics and morphological diversity relevant to Southeast Asia. The results demonstrate the feasibility of establishing scalable and interpretable datasets that integrate computational methods with expert knowledge. The proposed dataset serves as a robust resource for advancing hematology research and contributes to bridging traditional diagnostics with AI-driven clinical support systems.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470543/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research Progress on Color Image Quality Assessment.","authors":"Minjuan Gao, Chenye Song, Qiaorong Zhang, Xuande Zhang, Yankang Li, Fujiang Yuan","doi":"10.3390/jimaging11090307","DOIUrl":"10.3390/jimaging11090307","url":null,"abstract":"<p><p>Image quality assessment (IQA) aims to measure the consistency between an objective algorithm output and a subjective perception measurement. This article focuses on this complex relationship in the context of color image scenarios-color image quality assessment (CIQA). This review systematically investigates CIQA applications in image compression, processing optimization, and domain-specific scenarios, analyzes benchmark datasets and assessment metrics, and categorizes CIQA algorithms into full-reference (FR), reduced-reference (RR) and no-reference (NR) methods. In this study, color images are evaluated using a newly developed CIQA framework. Focusing on FR and NR methods, FR methods leverage reference images with machine learning, visual perception models, and mathematical frameworks, while NR methods utilize distortion-only features through feature fusion and extraction techniques. Specialized CIQA algorithms are developed for robotics, low-light, and underwater imaging. Despite progress, challenges remain in cross-domain adaptability, generalization, and contextualized assessment. Future directions may include prototype-based cross-domain adaptation, fidelity-structure balancing, spatiotemporal consistency integration, and CIQA-restoration synergy to meet emerging demands.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470951/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unsupervised Optical Mark Recognition on Answer Sheets for Massive Printed Multiple-Choice Tests.","authors":"Yahir Hernández-Mier, Marco Aurelio Nuño-Maganda, Said Polanco-Martagón, Guadalupe Acosta-Villarreal, Rubén Posada-Gómez","doi":"10.3390/jimaging11090308","DOIUrl":"10.3390/jimaging11090308","url":null,"abstract":"<p><p>The large-scale evaluation of multiple-choice tests is a challenging task from the perspective of image processing. A typical instrument is a multiple-choice question test that employs an answer sheet with circles or squares. Once students have finished the test, the answer sheets are digitized and sent to a processing center for scoring. Operators compute each exam score manually, but this task requires considerable time. While it is true that mature algorithms exist for detecting circles under controlled conditions, they may fail in real-life applications, even when using controlled conditions for image acquisition of the answer sheets. This paper proposes a desktop application for optical mark recognition (OMR) on the scanned multiple-choice question (MCQ) test answer sheets. First, we compiled a set of answer sheet images corresponding to 6029 exams (totaling 564,040 four-option answers) applied in 2024 in Tamaulipas, Mexico. Subsequently, we developed an image-processing module that extracts answers from the answer sheets and an interface for operators to perform analysis by selecting the folder containing the exams and generating results in a tabulated format. We evaluated the image-processing module, achieving a percentage of 96.15% of exams graded without error and 99.95% of 4-option answers classified correctly. We obtained these percentages by comparing the answers generated through our system with those generated by human operators, who took an average of 2 min to produce the answers for a single answer sheet, while the automated version took an average of 1.04 s.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470569/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficient Retinal Vessel Segmentation with 78K Parameters.","authors":"Zhigao Zeng, Jiakai Liu, Xianming Huang, Kaixi Luo, Xinpan Yuan, Yanhui Zhu","doi":"10.3390/jimaging11090306","DOIUrl":"10.3390/jimaging11090306","url":null,"abstract":"<p><p>Retinal vessel segmentation is critical for early diagnosis of diabetic retinopathy, yet existing deep models often compromise accuracy for complexity. We propose DSAE-Net, a lightweight dual-stage network that addresses this challenge by (1) introducing a Parameterized Cascaded W-shaped Architecture enabling progressive feature refinement with only 1% of the parameters of a standard U-Net; (2) designing a novel Skeleton Distance Loss (SDL) that overcomes boundary loss limitations by leveraging vessel skeletons to handle severe class imbalance; (3) developing a Cross-modal Fusion Attention (CMFA) module combining group convolutions and dynamic weighting to effectively expand receptive fields; and (4) proposing Coordinate Attention Gates (CAGs) to optimize skip connections via directional feature reweighting. Evaluated extensively on DRIVE, CHASE_DB1, HRF, and STARE datasets, DSAE-Net significantly reduces computational complexity while outperforming state-of-the-art lightweight models in segmentation accuracy. Its efficiency and robustness make DSAE-Net particularly suitable for real-time diagnostics in resource-constrained clinical settings.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470885/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"FDMNet: A Multi-Task Network for Joint Detection and Segmentation of Three Fish Diseases.","authors":"Zhuofu Liu, Zigan Yan, Gaohan Li","doi":"10.3390/jimaging11090305","DOIUrl":"10.3390/jimaging11090305","url":null,"abstract":"<p><p>Fish diseases are one of the primary causes of economic losses in aquaculture. Existing deep learning models have progressed in fish disease detection and lesion segmentation. However, many models still have limitations, such as detecting only a single type of fish disease or completing only a single task within fish disease detection. To address these limitations, we propose FDMNet, a multi-task learning network. Built upon the YOLOv8 framework, the network incorporates a semantic segmentation branch with a multi-scale perception mechanism. FDMNet performs detection and segmentation simultaneously. The detection and segmentation branches use the C2DF dynamic feature fusion module to address information loss during local feature fusion across scales. Additionally, we use uncertainty-based loss weighting together with PCGrad to mitigate conflicting gradients between tasks, improving the stability and overall performance of FDMNet. On a self-built image dataset containing three common fish diseases, FDMNet achieved 97.0% mAP50 for the detection task and 85.7% mIoU for the segmentation task. Relative to the multi-task YOLO-FD baseline, FDMNet's detection mAP50 improved by 2.5% and its segmentation mIoU by 5.4%. On the dataset constructed in this study, FDMNet achieved competitive accuracy in both detection and segmentation. These results suggest potential practical utility.</p>","PeriodicalId":37035,"journal":{"name":"Journal of Imaging","volume":"11 9","pages":""},"PeriodicalIF":2.7,"publicationDate":"2025-09-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12470551/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145150582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}