Object detection and segmentation using DenseNets and SIFT Keypoint Match

Xiwen Cui; Dongjun Huang; Wei Lei; Ammar Oad

doi:10.1117/12.2524255

6 May 2019 Object detection and segmentation using DenseNets and SIFT Keypoint Match

Xiwen Cui, Dongjun Huang, Wei Lei, Ammar Oad

Proceedings Volume 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018); 110692F (2019) https://doi.org/10.1117/12.2524255
Event: Tenth International Conference on Graphic and Image Processing (ICGIP 2018), 2018, Chengdu, China

Abstract

In this paper, we propose a model based on DenseNets and scale invariant feature transform (SIFT) keypoint match technology for object detection and segmentation. DenseNets is built on Convolutional Neural Networks (CNNs) with dense connections and used for semantic image segmentation. Our main idea is that, on the basis of the DenseNets model, we conduct the morphological processing, and apply the SIFT keypoint match technology to detect the object pixels. Opening operation and closing operation are the basic operations of morphological processing. They all consist of erosion operation and dilation operation but the order is different between them. The morphological processing combines two kinds of operations and can form a morphological filter which can filter the noise. The SIFT keypoint match algorithm is widely used to extract the invariant of position, scale and rotation so we use it to eliminate the misjudgment. Our experiments show that our method can acquire more accurate results compared with DenseNets model.

Citation Download Citation

Xiwen Cui, Dongjun Huang, Wei Lei, and Ammar Oad "Object detection and segmentation using DenseNets and SIFT Keypoint Match", Proc. SPIE 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018), 110692F (6 May 2019); https://doi.org/10.1117/12.2524255

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available