Share Email Print

Proceedings Paper

The method for detecting small lesions in medical image based on sliding window
Author(s): Guilai Han; Yuan Jiao
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

At present, the research on computer-aided diagnosis includes the sample image segmentation, extracting visual features, generating the classification model by learning, and according to the model generated to classify and judge the inspected images. However, this method has a large scale of calculation and speed is slow. And because medical images are usually low contrast, when the traditional image segmentation method is applied to the medical image, there is a complete failure. As soon as possible to find the region of interest, improve detection speed, this topic attempts to introduce the current popular visual attention model into small lesions detection. However, Itti model is mainly for natural images. But the effect is not ideal when it is used to medical images which usually are gray images. Especially in the early stages of some cancers, the focus of a disease in the whole image is not the most significant region and sometimes is very difficult to be found. But these lesions are prominent in the local areas. This paper proposes a visual attention mechanism based on sliding window, and use sliding window to calculate the significance of a local area. Combined with the characteristics of the lesion, select the features of gray, entropy, corner and edge to generate a saliency map. Then the significant region is segmented and distinguished. This method reduces the difficulty of image segmentation, and improves the detection accuracy of small lesions, and it has great significance to early discovery, early diagnosis and treatment of cancers.

Paper Details

Date Published: 7 October 2016
PDF: 5 pages
Proc. SPIE 9902, Fourth International Conference on Wireless and Optical Communications, 99020K (7 October 2016); doi: 10.1117/12.2262277
Show Author Affiliations
Guilai Han, Hainan Medical College (China)
Yuan Jiao, Hainan Medical College (China)

Published in SPIE Proceedings Vol. 9902:
Fourth International Conference on Wireless and Optical Communications
Maode Ma, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?