Design of Statistical Measures for the Assessment of Image Segmentation Schemes

Abstract

Image segmentation is discussed for years in numerous papers, but assessing its quality is mainly dealt with in recent works. Quality assessment is a primary concern for anyone working towards better segmentation tools. It both helps to objectively improve segmentation techniques and to compare performances with respect to other similar algorithms. In this paper we use a statistical framework to propose statistical measures capable to describe the performances of a segmentation scheme. All the measures rely on a ground-truth segmentation map that is supposed to be known and that serves as a reference when qualifying the results of any segmentation tool. We derive the analytical expression of several transition probabilities and show how to calculate them. An important conclusion from our study, often overlooked, is that performances can be content dependent, which means that one should adapt a measure to the content of an image.

Image segmentation evaluation: A survey of unsupervised methods

“…discrepancy methods evaluated segmented images based on the number of misclassified pixels versus the reference image, with penalties weighted proportional to the distance to the closest correctly classified pixel for that region [10][11][12][13][14]. Another group of discrepancy methods are based on the differences in the feature values measured from the segmented images and the reference image [15][16][17][18][19][20]. These methods have been extended to accommodate the problem when the number of objects differs between the segmented and reference images [21][22][23][24][25].…”

“…19), to describe the inter-region disparity. C is the sum of the per-region contrast measures, weighted by a function approximating the human contrast sensitivity curve.…”

Прямая Оценка Качества Программных Продуктов. Критерии И Тестовые Материалы

“…Еще одна мера оценки качества сегментации, основанная на подсчете неправильно классифицирован-ных пикселей и использующая байесовский подход, была предложена в [11]. В этой работе для случая одного сегмента вычисляются оценки вероятности того, что случайно выбранный пиксель на отсегмен-тированном изображении принадлежит сегменту и, соответственно, фону.…”

New benchmark for image segmentation evaluation

“…[11][12][13][14][15][25][26][27][28][29][30][31][32] Since the construction of the ground-truth segmentation for many real images is labor-intensive and sometimes not well or uniquely defined, most prior image segmentation methods are only tested on ͑1͒ some special classes of images used in special applications where the ground-truth segmentations are uniquely defined, ͑2͒ synthetic images where ground-truth segmentation is also well defined, and/or ͑3͒ a small set of real images.…”