Comparison of HOG, MSER, SIFT, FAST, LBP and CANNY Features for Cell Detection in Histopathological Images

Author Name(s): Şaban Öztürk, Bayram Akdemir
Author Email:


Cell segmentation and counting has a very important role in diagnosing diseases and in the treatment process. But the complexity of the histopathological images and the differences in cell groups make this process very difficult, even for an expert. In order to facilitate this process, analysis of histopathological images is performed by using computer vision methods. This paper presents the use of different feature extraction methods for cell detection in histopathological images and the comparison of the results of these algorithms. For this reason, HOG, MSER, SIFT, FAST, LBP and CANNY feature extraction algorithms are used. The aim of the study is to determine cells using different feature extraction methods and to determine which of these feature extraction algorithms will be more successful. Firstly, image pre-processing has been applied to clear the noises in the histopathological images. Then, feature extraction algorithms are applied to image, respectively. Finally, the successes of different feature extraction algorithms have been compared.


HOG, MSER, SIFT, FAST, LBP, canny, histopathological image, cell counting.


Computer aided diagnosis studies in biomedical engineering field are very useful for human health and diagnosis of diseases. In particular, the state of cellular structures and cell forms present useful information about many diseases and cancer types [1]. For this reason, cell forms and behaviors in histopathological images have become one of the main research areas. However, because of the negative and compelling properties of these images, such as unwanted light effects, low contrast, differences in cell shapes etc. analysis takes a long time. Again for this reason, even if these images are evaluated by an expert, histopathological image analysis is quite challenging and time-consuming. Also, it requires skill and experience [2]. In order to overcome all these difficulties, a computer-assisted automatic cell detection system is needed. Image processing algorithms which are used heavily in these systems and directly affect the success of the system have an important effect in this context [3, 4]. Some studies in the literature for cell detection from histopathological images are; compact Hough and radial map [5], logistic regression classifier [6], watershed segmentation [7], mean-shift clustering based method [8]. In the image segmentation or classification process using image processing techniques, features obtained from images are affecting success. For this reason different features are used for different problems. The characteristics obtained by different feature extraction methods applied on the same problem can lead to different success rates. For this reason, many feature extraction methods have been proposed in the literature. For example, feature extraction methods such as SIFT and SURF are used for interest region detection, and LBP and LPQ are preferred for texture classification operations [9]. In this study, the success of well-known feature extraction methods for cell detection in histopathological images is investigated. For this purpose, HOG, MSER, SIFT, FAST, LBP feature extraction methods and canny edge detection algorithm are used. In order to detect the cells through the features obtained from the histopathological image using the mentioned algorithms, the feature sets need only cover the cells. For this reason, when applied to raw images, they fail because of disturbing factors such as noise. In this study, firstly RGB color space to HSV color space conversion is done. Then, noise in the V component of the HSV color space is cleaned using wiener filter [10]. Then, image sharpening is applied to the image. With the help of pre-processing, feature extraction algorithms have increased the chances of catching only the cells. In the next step, HOG [11], MSER [12], SIFT [13], FAST [14], LBP [15] and canny [16] algorithms are applied respectively. Finally, cells are identified with the help of extracted features and the success of each algorithm is compared with other algorithms


In this study, the success of well-known feature extraction methods is investigated for cell detection in histopathological images. It has been understood that feature extraction algorithms cannot achieve the desired success on raw images. For this reason, preprocessing is applied before image properties are extracted. Then feature extraction algorithms are applied to the cleaned image. Each feature extraction algorithm operates on its own characteristic feature. In order for the comparison process to be objective, these characteristics are preserved and applied in the same conditions. At the end of the comparison process, MSER, LBP and canny algorithms detect cells with high success. However, the LBP and canny algorithms incorrectly mark many regions. In this case, the accuracy rate is low. The HOG algorithm is more advantageous than the LBP and canny algorithms because it has a low error rate when it has high success in determining the cells. In line with this information, the MSER algorithm can be used as the first option and the HOG algorithm can be used as the second option. But, the selection process may vary depending on the application.

992 total views, 1 views today

Download PDF File

About the author: admin