The sonogram is currently an effective cancer screening and diagnosis way due to the convenience and harmlessness in humans. Traditionally, lesion boundary segmentation is first adopted and then classification is conducted, to reach the judgment of benign or malignant tumor. In addition, sonograms often contain much speckle noise and intensity inhomogeneity. This study proposes a novel benign or malignant tumor classification system, which comprises intensity inhomogeneity correction and stacked denoising autoencoder (SDAE), and it is suitable for small-size dataset. A classifier is established by extracting features in the multilayer training of SDAE; automatic analysis of imaging features by the deep learning algorithm is applied on image classification, thus allowing the system to have high efficiency and robust distinguishing. In this study, two kinds of dataset (private data and public data) are used for deep learning models training. For each dataset, two groups of test images are compared: the original images and the images after intensity inhomogeneity correction, respectively. The results show that when deep learning algorithm is applied on the sonograms after intensity inhomogeneity correction, there is a significant increase of the tumor distinguishing accuracy. This study demonstrated that it is important to use preprocessing to highlight the image features and further give these features for deep learning models. In this way, the classification accuracy will be better to just use the original images for deep learning.

Ultrasonic detection is currently an effective cancer screening and diagnosis method due to the convenience and harmlessness to human. A set of systems are investigated in this article to pick up the complete tumor outline. After noises in ultrasonic tumor images are removed automatically and areas of different characteristics are distinguished by cutting tumor outlines, images with similar attributes are integrated. Finally the tumor outline is described precisely to facilitate the computer tumor classification.Because ultrasound images often contain a lot of noises, preprocessing removes spot noises by Gaussian filter and select then the appropriate threshold to binarize images. ROD (Rank-ordered Differences) method is applied to calculate the grey level difference between neighbour pixels and the particular pixel to detect pixels contaminated by noises. Images become converged by interactive iteration of two masks of different sizes and a false boundary is obtained after Sobel treatment.ut the original image into small regions by watershed conversion, label regions and calculate the standard deviation within a region. If the standard deviation is close to the region with the false boundary, the region is considered to be the tumor region.