COLOR RECOGNITION METHODS IN DIGITAL IMAGES: A REVIEW OF CLASSICAL AND DEEP LEARNING APPROACHES

Computer Engineering

Color detection is a fundamental computer vision task essential for applications spanning object recognition, robotics, image retrieval, and accessibility. Its wide applicability necessitates accurate, efficient, and robust methods capable of identifying colors under diverse conditions, including lighting variations and image complexity. This survey aims to comprehensively review and analyze techniques for color detection in digital images, emphasizing implementations utilizing Python and prevalent computer vision libraries like OpenCV. We examine numerous approaches, from foundational methods involving color space analysis (RGB, HSV, LAB) and thresholding techniques (global, adaptive) to more advanced strategies employing clustering algorithms (K-Means, Fuzzy C-Means). The review further covers the use of standard machine learning classifiers, such as K-Nearest Neighbors (KNN), and the increasing adoption of deep learning models, including Convolutional Neural Networks (CNNs), often combined with feature extraction based on color histograms, contrast, and saturation. The literature indicates that methods like HSV thresholding and KNN/K-Means clustering provide effective solutions for many common color identification tasks. Concurrently, deep learning approaches demonstrate significant potential for addressing challenging scenarios, particularly regarding lighting invariance, albeit with higher computational demands. This survey seeks to guide researchers in selecting suitable methodologies by comparing various techniques and highlights areas for future research, such as real-time implementation and enhanced robustness for complex visual environments. The summarized algorithms offer valuable tools for integration into software applications requiring automated color analysis.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.