Image segmentation is one of the most fundamental topics in image processing. Although there are many segmentation studies in literature, approaches that combine different image processing techniques have come to the forefront today. This study proposes a new approach based on color quantization, edge detection, and region growing. In the first stage of this three-stage approach, color quantization is performed. Equally spaced multiple threshold selection is performed on each color channel. In the second stage, edge detection is performed on the quantized color image. Thus, regions consisting of pixels without edges are obtained. These regions are expanded with the help of region growing to ensure that the most similar pixels are together. The success of the approach is tested on the Weizmann single-object image dataset containing 100 color images. Precision, recall, and F-score results obtained in MATLAB were compared with different thresholds used in the quantization phase. As a result, it was determined that the increase in the number of colors decreased the segmentation success. In the future, studies that focus on the detailed selection of pixels of the object will be carried out using different quantization and edge detection methods.