International Journal of Image Processing and Smart Sensors

Edge Detection in Image Processing using Sobel Operator

Purva Kamat Mhamai, Niharika M., Vidya H. R., Chandana M. — Thu, 29 Jan 2026 00:00:00 +0000

Biomedical applications of nanoparticles are rapidly expanding, supporting diagnostics, therapeutics, biosensing, implant surface improvements, and advanced medical textiles. Medical imaging benefits from image enhancement, segmentation, pattern recognition, and computer-aided diagnosis (CAD) systems. Edge detection plays a key role in identifying structures in medical images. A CAD system is developed to classify abnormal cervical cells using morphological and statistical edge-based segmentation, achieving 98% accuracy. Because image processing on large datasets is computationally intensive, parallel and multithreaded Sobel edge detection algorithms help reduce processing time. Quantum image processing further accelerates edge detection by using quantum superposition and parallelism with improved Sobel operators. Several improved Sobel-based methods enhance edge clarity, reduce noise, and extract detailed contours. Fractional order derivative (FOD)-based Sobel performs better than Prewitt and Laplacian in PSNR, SSIM, and FSIM metrics for fish images. Ant Colony Optimization (ACO)-based edge detection accelerates convergence using modified pheromone updates. Detection, medical diagnosis, SEM image analysis, and contrast enhancement tasks also employ edge detection and morphological operations. The system is implemented in a 512×512 image that takes 0.009ms in the processing system. For a 3×3-pixel block, the design uses 22 logic gates with a minimum delay of 1.5 fs. It occupies an area of 111 nm² and operates at a supply voltage of 1.05V. The circuit’s average power consumption is measured at 2.27 μWat.

An Edge-driven Framework for Barcode Localization and Recognition Using OpenCV

Simranjeet Kaur, Ameya K. Naik, Kiran V. Ajetrao — Thu, 26 Feb 2026 00:00:00 +0000

Barcodes have become a part of our lives, from scanning groceries at the supermarket to buying apparel at the mall. Barcodes are used extensively by all businesses, whether retail or production lines at a factory. Thus, it becomes necessary to scan all the barcodes efficiently with the barcode scanner. Scanning these barcodes seems like a cakewalk, but there is a lot of work behind the scenes. Barcodes may not be scanned properly if they are blurred, not well-lit or if the image of the barcode is rotated. This study has taken into account such barcodes with some environmental noise and applied traditional edge-based processing with Pyzbar. The traditional edge detection techniques that were applied on the barcode images are Canny, Sobel, Scharr and Prewitt. These techniques were applied to the Medium Barcode 1D dataset. The dataset is an open dataset. On applying the edge-based techniques, it was found that Scharr obtained the highest F1-score of 0.697 with a success rate of 76.15%. Though various machine learning and deep learning methods are present, they are computationally expensive. Thus, traditional edge-based processing must be evaluated as they consume less computing resources and memory.

MATLAB-based Performance Assessment of a GNRFET-based Ion Sensitive Field Effect Transistor for Point-of-Care Diagnosis and Precision Medicine

Muhammad Johirul Islam, Iqbal Bahar Chowdhury — Mon, 23 Mar 2026 00:00:00 +0000

In this work, an ion-sensitive field-effect transistor (ISFET) based on graphene nanoribbon (GNR) channel FET (GNRFET) geometry has been explored as an aggressively-scaled architecture for pH-based wearable sweat sensors and precision medicine with improved performance metrics. The proposed sensor has been implemented in a MATLAB simulation framework interlinking the SPICE models published in the literature for the GNRFET and the electrochemical stages required to build the ISFET structure. The ISFET is biased in the weak inversion regime through a transconductance amplifier configuration to ensure minimum power consumption and higher current sensitivity. The proposed ISFET-based pH sensor exhibits a moderate voltage sensitivity of 48.3 mV/pH, a current sensitivity of 0.45 dec/pH and an ultra-low power consumption of only 6 pW using an extremely scaled device active area of only 15 nm × 0.46 nm. Therefore, the proposed ISFET device has excellent promise in the realization of ultra-scaled pH-based wearable sweat sensors and DNA-sensors with higher sensitivity, and lower power consumption as demanded by the point-of-care (POC) diagnosis, and required by the DNA sequencing in the precision medicine, both for modern health care.

Designing a Communication-efficient and Privacy-preserving Collaborative Image Processing Framework for Distributed Smart Sensor Networks under Bandwidth and Energy Constraints

Md. Ali — Sat, 02 May 2026 00:00:00 +0000

This work investigates the challenges associated with distributed smart sensor networks for large-scale image acquisition and analysis, particularly in applications such as environmental monitoring, smart cities, and surveillance. Despite their increasing deployment, these systems face significant constraints, including limited communication bandwidth, restricted energy resources, and critical privacy concerns due to the transmission of raw image data to centralized servers. To address these challenges, this work proposes a novel communication-efficient and privacy-preserving collaborative image processing framework. The proposed approach leverages edge-based feature extraction to process images locally at sensor nodes, thereby reducing the need to transmit high-volume raw data. In addition, adaptive compression techniques are employed to further minimize communication costs without substantially degrading the quality of extracted features. The framework also incorporates federated learning, enabling distributed model training across multiple nodes while keeping sensitive data localized. To enhance security, secure aggregation mechanisms are integrated to ensure that shared model updates remains confidential and resistant to potential adversarial attacks. Experimental results demonstrate that the proposed framework achieves significant performance improvements, including up to a 65% reduction in communication overhead and up to a 40% decrease in energy consumption compared to conventional centralized approaches. Importantly, these efficiency gains are achieved while maintaining competitive accuracy in image processing and analysis tasks. Overall, this work presents a scalable, energy-efficient, and privacy-aware solution for next-generation distributed sensing systems, making it highly suitable for deployment in resource-constrained and privacy-sensitive environments.

A Kalman Filter-Enhanced Image Reconstruction Technique for Privacy-Preserving Human Action Recognition Using mmWave Radar Smart Sensors

Belay Goshu — Wed, 10 Jun 2026 00:00:00 +0000

Millimeter-Wave (mmWave) radar provides a privacy-preserving alternative to RGB cameras for Human Action Recognition (HAR). However, its sparse and noisy point clouds often limit classification accuracy. Existing temporal filtering methods typically apply Kalman filters only to centroid tracking, rather than full image reconstruction. This paper proposes a Kalman filter-enhanced image reconstruction framework that integrates a full state-space model into the sparse-to-dense heatmap generation pipeline. Raw point clouds are first preprocessed using adaptive DBSCAN clustering. A Kalman filter is then employed to estimate a dense 64×64 heatmap based on constant-velocity dynamics. To efficiently manage the high-dimensional 4096-state vector, the study assume pixel-wise independence (a diagonal covariance matrix), reducing computational complexity from O (N³) to O(N) per update. The filtered points are subsequently projected onto a 2D grid using Gaussian kernel density estimation. Finally, a lightweight CNN-LSTM network performs action classification. Evaluated on the RadHAR dataset (five activities, ten subjects) using a Texas Instruments IWR1443 radar and an NVIDIA Jetson Nano, the proposed method achieves 94.78% accuracy. This represents a substantial improvement of 13.43 percentage points over vanilla reconstruction (81.35%) and 5.05 percentage points over centroid-only Kalman filtering (89.73%). The approach also demonstrates strong robustness at low frame rates (86.37% at 5 fps vs. 44.38%) and under occlusions (SSIM of 0.732 vs. 0.418). Privacy evaluations confirm zero successful face recognition and only chance-level re-identification performance (8.2%). Notably, the Kalman filter executes in just 12.8 ms per frame on the Jetson Nano, enabling real-time inference at up to 19 fps. By integrating a full state-space Kalman filter into the mmWave radar heatmap reconstruction process, the proposed framework significantly enhances accuracy, occlusion resilience, and temporal coherence, while maintaining strong privacy protection and satisfying strict edge-device real-time constraints.