PERFORMANCE BENCHMARKING IN COLOUR MODEL Article Swipe
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.36948/ijfmr.2025.v07i02.37944
· OA: W4410029913
This study of different colour models in Image processing aims for developing a comprehensive model that can be used for comparing the vast range of models. The following comparison will be evaluated on the basis of some critical factors mainly listed as: accuracy, processing time, usage of space, graphical processing unit(GPU) performance, central processing unit(CPU) performance and some other relevant metrics. Many research work on colour models have been done mathematically like a comprehensive study on different available colour models for suitable computer vision tasks. The core intent of this paper is to analyse the colour models performance on the basis of the utilisation of a customised dataset. And perform conversion of each image in all the major colour models, followed by the conversion, we have employed convolutional neural network (CNN) for object detection tasks, supported by edge detection using morphological filtering with the operators like dilation and erosion for an enhanced and smoother detection. We have used YOLO V5 and Mediapipe from google for a lightweight model (key findings) .The significance of this project helps to identify the effective colour model without having to test out every model for a required task. Due to this comprehensive comparative table / survey one can choose the model efficiently for their required colour spaces and object detection task.
