Learning Strengths and Weaknesses of Classifiers for RGB-D Semantic Segmentation

سال انتشار: 1394
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 507

متن کامل این مقاله منتشر نشده است و فقط به صورت چکیده یا چکیده مبسوط در پایگاه موجود می باشد.
توضیح: معمولا کلیه مقالاتی که کمتر از ۵ صفحه باشند در پایگاه سیویلیکا اصل مقاله (فول تکست) محسوب نمی شوند و فقط کاربران عضو بدون کسر اعتبار می توانند فایل آنها را دریافت نمایند.

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

ICMVIP09_072

تاریخ نمایه سازی: 6 اسفند 1395

چکیده مقاله:

3D scene understanding is an open challenge in the field of computer vision. Most of the focus is on 2D methods in which the semantic labeling of each RGB pixel is considered. But, in this paper, the 3D semantic labeling of RGB-D images is considered. In the proposed method, to extract some meaningful features, the superpixel generation algorithm is applied to the RGB image to segment it into a set of disjoint pixels. After that, the set of three powerful classifiers are utilized to semanticallylabel each superpixel. In the proposed method, the probability outputs of these classifiers are concatenated as the novel feature vector for each superpixel. Consequently, to analyze the strengthsand weaknesses of each classifier, the conditional random field framework is used to improve the contextual relationships among neighboring superpixels. The unary potential function of the conditional random field is learned based on these new feature vectors. The proposed method is evaluated on the challenging NYU-V2 RGB-D dataset and improves the pixel average accuracy compared to previous methods.

نویسندگان

Fahimeh Fooladgar

Department of Computer Engineering Sharif University of Technology Tehran, Iran