DESIGN AND IMPLEMENTATION OF A LOW-POWER FPGA ACCELERATOR FOR REAL-TIME SKIN CANCER DETECTION USING OPTIMIZED IMAGE PROCESSING TECHNIQUES

Abstract

Skin cancer is one of the most prevalent and life-threatening diseases worldwide, where early diagnosis plays a crucial role in improving patient survival and treatment outcomes. Traditional software-based skin lesion analysis systems often suffer from high computational complexity, increased processing time, and significant power consumption, limiting their applicability in real-time medical environments. This paper presents a low-power FPGA-based real-time skin cancer detection system that integrates advanced image processing algorithms with pipeline balancing and parallelism optimization techniques. The proposed methodology consists of image acquisition, preprocessing, lesion segmentation, feature extraction, feature fusion, and classification stages. Preprocessing techniques are employed to enhance image quality and remove noise, while segmentation algorithms isolate the suspicious lesion region from surrounding healthy skin. Subsequently, discriminative features related to color, texture, shape, and structural characteristics are extracted using image processing methods such as Gray Level Co-occurrence Matrix (GLCM), Discrete Wavelet Transform (DWT), and statistical analysis. These features are combined and supplied to a classification module for identifying skin lesions as normal, benign, or malignant. To achieve highspeed real-time performance, the complete processing chain is implemented on FPGA hardware using Verilog HDL with optimized pipeline balancing and parallel execution of computational tasks. The proposed architecture significantly reduces processing latency, improves throughput, and minimizes power consumption while maintaining high diagnostic accuracy. Experimental results demonstrate that the FPGAbased implementation provides an efficient, scalable, and portable solution for computer-aided dermatological diagnosis, making it suitable for embedded healthcare systems and realtime clinical applications.