Electronics

Research

14 pages, 1588 KB

Open AccessArticle

Zero-Shot SAM for Food Image Segmentation

by Saeed S. Alahmari, Michael R. Gardner and Tawfiq Salem

Electronics 2025, 14(21), 4316; https://doi.org/10.3390/electronics14214316 - 3 Nov 2025

Viewed by 734

Recent advances in foundation models have enabled strong zero-shot performance across vision tasks, yet their effectiveness for fine-grained domains such as food image segmentation remains underexplored. This study investigates the zero-shot capabilities of the Segment Anything Model (SAM) for segmenting food images. We [...] Read more.

Recent advances in foundation models have enabled strong zero-shot performance across vision tasks, yet their effectiveness for fine-grained domains such as food image segmentation remains underexplored. This study investigates the zero-shot capabilities of the Segment Anything Model (SAM) for segmenting food images. We evaluate two prompting strategies—bounding boxes and coordinate points—and compare SAM’s results against established deep learning baselines, U-Net and DeepLabv3+, both trained on the FoodBin-17k dataset with pixel-wise binary annotations. Using bounding box prompts, zero-shot SAM achieved a Dice coefficient of 0.790, closely matching U-Net (0.759), while DeepLabv3+ attained 0.918. Coordinate-based prompting yielded comparable results, with the optimal configuration (75 coordinate points) also reaching a Dice coefficient of 0.790. These results highlight SAM’s strong potential for generalizing to novel visual categories without task-specific retraining. This study further provides insights into how prompt type influences segmentation quality, offering practical guidance for applying SAM in food-related and other fine-grained segmentation applications. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

19 pages, 2911 KB

Open AccessArticle

MCFI-Net: Multi-Scale Cross-Layer Feature Interaction Network for Landslide Segmentation in Remote Sensing Imagery

by Jianping Liao and Lihua Ye

Electronics 2025, 14(21), 4190; https://doi.org/10.3390/electronics14214190 - 27 Oct 2025

Viewed by 344

Abstract

Accurate and reliable detection of landslides plays a crucial role in disaster prevention and mitigation efforts. However, due to unfavorable environmental conditions, uneven surface structures, and other disturbances similar to those of landslides, traditional methods often fail to achieve the desired results. To [...] Read more.

Accurate and reliable detection of landslides plays a crucial role in disaster prevention and mitigation efforts. However, due to unfavorable environmental conditions, uneven surface structures, and other disturbances similar to those of landslides, traditional methods often fail to achieve the desired results. To address these challenges, this study introduces a novel multi-scale cross-layer feature interaction network, specifically designed for landslide segmentation in remote sensing images. In the MCFI-Net framework, we adopt the encoder–decoder as the foundational architecture, and integrate cross-layer feature information to capture fine-grained local textures and broader contextual patterns. Then, we introduce the receptive field block (RFB) into the skip connections to effectively aggregate multi-scale contextual information. Additionally, we design the multi-branch dynamic convolution block (MDCB), which possesses both dynamic perception ability and multi-scale feature representation capability. The comprehensive evaluation conducted on both the Landslide4Sense and Bijie datasets demonstrates the superior performance of MCFI-Net in landslide segmentation tasks. Specifically, on the Landslide4Sense dataset, MCFI-Net achieved a Dice score of 0.7254, a Matthews correlation coefficient (Mcc) of 0.7138, and a Jaccard score of 0.5699. Similarly, on the Bijie dataset, MCFI-Net maintained high accuracy with a Dice score of 0.8201, an Mcc of 0.8004, and a Jaccard score of 0.6951. Furthermore, when evaluated on the optical remote sensing dataset EORSSD, MCFI-Net obtained a Dice score of 0.7770, an Mcc of 0.7732, and a Jaccard score of 0.6571. Finally, ablation experiments carried out on the Landslide4Sense dataset further validated the effectiveness of each proposed module. These results affirm MCFI-Net’s capability in accurately identifying landslide regions from complex remote sensing imagery, and it provides great potential for the analysis of geological disasters in the real world. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

11 pages, 2091 KB

Open AccessArticle

Underwater Image Enhancement Method Based on Vision Mamba

by Yongjun Wang, Zhuo Chen, Maged Al-Barashi and Zeyu Tang

Electronics 2025, 14(17), 3411; https://doi.org/10.3390/electronics14173411 - 27 Aug 2025

Viewed by 808

Abstract

To address issues like haze, blurring, and color distortion in underwater images, this paper proposes a novel underwater image enhancement model called U-Vision Mamba, built on the Vision Mamba framework. The core innovation lies in a U-shaped network encoder for multi-scale feature extraction, [...] Read more.

To address issues like haze, blurring, and color distortion in underwater images, this paper proposes a novel underwater image enhancement model called U-Vision Mamba, built on the Vision Mamba framework. The core innovation lies in a U-shaped network encoder for multi-scale feature extraction, combined with a novel multi-scale sparse attention fusion module to effectively aggregate these features. This fusion module leverages sparse attention to capture global context while preserving fine details. The decoder then refines these aggregated features to generate high-quality underwater images. Experimental results on the UIEB dataset demonstrate that U-Vision Mamba significantly reduces image blurring and corrects color distortion, achieving a PSNR of 25.65 dB and an SSIM of 0.972. Both comprehensive subjective evaluation and objective metrics confirm the model’s superior performance and robustness, making it a promising solution for improving the clarity and usability of underwater imagery in applications like marine exploration and environmental monitoring. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

15 pages, 1563 KB

Open AccessArticle

FFMN: Fast Fitting Mesh Network for Monocular 3D Human Reconstruction in Live-Line Work Scenarios

by Guokai Liang, Jie Zhou, Fan Yang, Guocheng Lin, Jiajian Luo, Xin Xie, Peng Zhang and Zhe Li

Electronics 2025, 14(12), 2362; https://doi.org/10.3390/electronics14122362 - 9 Jun 2025

Viewed by 987

Abstract

In live-line power distribution operations, 3D pose and action recognition of workers holds critical significance for safety assurance and intelligent monitoring. We propose a novel neural network architecture for fast fitting-based parametric 3D human reconstruction (FFMN) from monocular images in live-line work scenarios. [...] Read more.

In live-line power distribution operations, 3D pose and action recognition of workers holds critical significance for safety assurance and intelligent monitoring. We propose a novel neural network architecture for fast fitting-based parametric 3D human reconstruction (FFMN) from monocular images in live-line work scenarios. FFMN employs convolutional neural networks to extract feature information from input images and adopts an optimization strategy for inverse problems by reprojecting keypoints from the human model onto feature maps to acquire feedback. A transformer-based updater module then adjusts the model to better align with the person in the image. Unlike conventional regression or recurrent models, FFMN trains faster, utilizes fewer parameters, and achieves shorter inference times. Moreover, our FFMN demonstrates a significant inference speed advantage (latency = 15 ms) on the 3DPW and Human3.6M datasets while maintaining competitive accuracy (MPJPE < 50 mm), highlighting its high practicability in real-world applications. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

12 pages, 2532 KB

Open AccessArticle

Application of Deep Dilated Convolutional Neural Network for Non-Flat Rough Surface

by Chien-Ching Chiu, Yang-Han Lee, Wei Chien, Po-Hsiang Chen and Eng Hock Lim

Electronics 2025, 14(6), 1236; https://doi.org/10.3390/electronics14061236 - 20 Mar 2025

Viewed by 680

Abstract

In this paper, we propose a novel deep dilated convolutional neural network (DDCNN) architecture to reconstruct periodic rough surfaces, including their periodic length, dielectric constant, and shape. Historically, rough surface problems were addressed through optimization algorithms. However, these algorithms are computationally intensive, making [...] Read more.

In this paper, we propose a novel deep dilated convolutional neural network (DDCNN) architecture to reconstruct periodic rough surfaces, including their periodic length, dielectric constant, and shape. Historically, rough surface problems were addressed through optimization algorithms. However, these algorithms are computationally intensive, making the process very time-consuming. To resolve this issue, we provide measured scattered fields as training data for the DDCNN to reconstruct the periodic length, dielectric constant, and shape. The numerical results demonstrate that DDCNN can accurately reconstruct rough surface images under high noise levels. In addition, we also discuss the impacts of the periodic length and dielectric constant of the rough surface on the shape reconstruction. Notably, our method achieves excellent reconstruction results compared to DCNN even when the period and dielectric coefficient are unknown. Finally, it is worth mentioning that the trained network model completes the reconstruction process in less than one second, realizing efficient real-time imaging. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

20 pages, 8948 KB

Open AccessArticle

Detection of Sealing Surface of Electric Vehicle Electronic Water Pump Housings Based on Lightweight YOLOv8n

by Li Sun, Yi Shen, Jie Li, Weiyu Jiang, Xiang Bian and Mingxin Yuan

Electronics 2025, 14(2), 258; https://doi.org/10.3390/electronics14020258 - 9 Jan 2025

Viewed by 993

Abstract

Due to the characteristics of large size differences and shape variations in the sealing surface of electric vehicle electronic water pump housings, and the shortcomings of traditional YOLO defect detection models such as large volume and low accuracy, a lightweight defect detection algorithm [...] Read more.

Due to the characteristics of large size differences and shape variations in the sealing surface of electric vehicle electronic water pump housings, and the shortcomings of traditional YOLO defect detection models such as large volume and low accuracy, a lightweight defect detection algorithm based on YOLOv8n (You Only Look Once version 8n) is proposed for the sealing surface of electric vehicle electronic water pump housings. First, on the basis of introducing the MoblieNetv3 module, the YOLOv8n network structure is redesigned, which not only achieves network lightweighting but also improves the detection accuracy of the model. Then, DualConv (Dual Convolutional) convolution is introduced and the CMPDual (Cross Max Pooling Dual) module is designed to further optimize the detection model, which reduces redundant parameters and computational complexity of the model. Finally, in response to the characteristics of large size differences and shape variations in sealing surface defects, the Inner-WIoU (Inner-Wise-IoU) loss function is used instead of the CIoU (Complete-IoU) loss function in YOLOv8n, which improves the positioning accuracy of the defect area bounding box and further enhances the detection accuracy of the model. The ablation experiment based on the dataset constructed in this paper shows that compared with the YOLOv8n model, the weight of the proposed model is reduced by 61.9%, the computational complexity is reduced by 58.0%, the detection accuracy is improved by 9.4%, and the mAP@0.5 is improved by 6.9%. The comparison of detection results from different models shows that the proposed model has an average improvement of 6.9% in detection accuracy and an average improvement of 8.6% on mAP@0.5, which indicates that the proposed detection model effectively improves defect detection accuracy while ensuring model lightweighting. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

21 pages, 6186 KB

Open AccessArticle

Automatic Measurement of Comprehensive Skin Types Based on Image Processing and Deep Learning

by Jianghong Ran, Guolong Dong, Fan Yi, Li Li and Yue Wu

Electronics 2025, 14(1), 49; https://doi.org/10.3390/electronics14010049 - 26 Dec 2024

Viewed by 4283

Abstract

The skin serves as a physical and chemical barrier, effectively protecting us against the external environment. The Baumann Skin Type Indicator (BSTI) classifies skin into 16 types based on traits such as dry/oily (DO), sensitive/resistant (SR), pigmented/nonpigmented (PN), and wrinkle-prone/tight (WT). Traditional assessments [...] Read more.

The skin serves as a physical and chemical barrier, effectively protecting us against the external environment. The Baumann Skin Type Indicator (BSTI) classifies skin into 16 types based on traits such as dry/oily (DO), sensitive/resistant (SR), pigmented/nonpigmented (PN), and wrinkle-prone/tight (WT). Traditional assessments are time-consuming and challenging as they require the involvement of experts. While deep learning has been widely used in skin disease classification, its application in skin type classification, particularly using multimodal data, remains largely unexplored. To address this, we propose an improved Inception-v3 model incorporating transfer learning, based on the four-dimensional classification of the Baumann Skin Type Index (BSTI), which demonstrates outstanding accuracy. The dataset used in this study includes non-invasive physiological indicators, BSTI questionnaires, and skin images captured under various light sources. By comparing performance across different light sources, regions of interest (ROI), and baseline models, the improved Inception-v3 model achieved the best results, with accuracy reaching 91.11% in DO, 81.13% in SR, 91.72% in PN, and 74.9% in WT, demonstrating its effectiveness in skin type classification. This study surpasses traditional classification methods and previous similar research, offering a new, objective approach to measuring comprehensive skin types using multimodal and multi-light-source data. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures

Figure 1

13 pages, 4672 KB

Open AccessArticle

A Four-Point Orientation Method for Scene-to-Model Point Cloud Registration of Engine Blades

by Duanjiao Li, Ying Zhang, Ziran Jia, Zhiyu Wang, Qiu Fang and Xiaogang Zhang

Electronics 2024, 13(23), 4634; https://doi.org/10.3390/electronics13234634 - 25 Nov 2024

Cited by 1 | Viewed by 1313

Abstract

The use of 3D optical equipment for multi-view scanning is a promising approach to assessing the processing errors of engine blades. However, incomplete scanned point cloud data may impact the accuracy of point cloud registration (PCR). This paper proposes a four-point orientation point [...] Read more.

The use of 3D optical equipment for multi-view scanning is a promising approach to assessing the processing errors of engine blades. However, incomplete scanned point cloud data may impact the accuracy of point cloud registration (PCR). This paper proposes a four-point orientation point cloud registration method to improve the efficiency and accuracy of the coarse registration of turbine blades and prevent PCR failure. First, the point cloud is divided into four labeling blocks based on a principal component analysis. Second, keypoints are detected in each block based on their distance from the plane formed by the principal axes and described with a location-label descriptor based on their position. Third, a keypoint pair set is chosen based on the descriptor, and a suitable keypoint base is selected through singular value decomposition to obtain the final rigid transformation. To verify the effectiveness of the method, experiments are conducted on different blades. The results demonstrate the improved performance and efficiency of the proposed method of coarse registration for turbine blades. Full article

(This article belongs to the Special Issue Advanced Machine Learning Technologies and Their Applications in Intelligent Imaging and Image Processing)

► Show Figures