Full Record

New Search | Similar Records

Author
Title General Object Detection Using Superpixel Preprocessing
URL
Publication Date
Discipline/Department Computer Vision
University/Publisher Linköping University
Abstract The objective of this master’s thesis work is to evaluate the potential benefit of a superpixel preprocessing step for general object detection in a traffic environment. The various effects of different superpixel parameters on object detection performance, as well as the benefit of including depth information when generating the superpixels are investigated. In this work, three superpixel algorithms are implemented and compared, including a proposal for an improved version of the popular Spectral Linear Iterative Clustering superpixel algorithm (SLIC). The proposed improved algorithm utilises a coarse-to-fine approach which outperforms the original SLIC for high-resolution images. An object detection algorithm is also implemented and evaluated. The algorithm makes use of depth information obtained by a stereo camera to extract superpixels corresponding to foreground objects in the image. Hierarchical clustering is then applied, with the segments formed by the clustered superpixels indicating potential objects in the input image. The object detection algorithm managed to detect on average 58% of the objects present in the chosen dataset. It performed especially well for detecting pedestrians or other objects close to the car. Altering the density distribution of the superpixels in the image yielded an increase in detection rate, and could be achieved both with or without utilising depth information. It was also shown that the use of superpixels greatly reduces the amount of computations needed for the algorithm, indicating that a real-time implementation is feasible.
Subjects/Keywords superpixels; SLIC; coarse-to-fine; segmentation; general object detection; cityscapes; traffic; image processing; clustering; Computer Vision and Robotics (Autonomous Systems); Datorseende och robotik (autonoma system)
Language en
Country of Publication se
Record ID oai:DiVA.org:liu-140874
Repository diva
Date Indexed 2020-01-03

Sample Search Hits | Sample Images | Cited Works

…in [22]. Their algorithm, Superpixels Extracted via Energy-Driven Sampling (SEEDS), can achieve real-time performance through the use of a coarse-to-fine approach. The pixels are iteratively assigned to superpixels through the…

…In this thesis, the potential use of a coarse-to-fine update strategy confined to superpixel boundaries for use together with the straightforward SLIC algorithm is investigated. The algorithm is compared to an implementation of the basic SLIC…

…algorithm as well as a coarse-to-fine energy optimisation algorithm akin to [22] and [24]. 2.3 Superpixel Algorithms As part of this thesis, three superpixel algorithms are implemented and compared. 1. Spectral Linear Iterative…

…Clustering (SLIC) - A superpixel algorithm proposed by Achanta et.al. in [2]. It is popular due to its simplicity and performance. 2. Coarse-To-Fine Superpixel Segmentation (CTF) - A modified version the coarse-to-fine block…

…update algorithms presented in [22, 24]. 3. Coarse-To-Fine Spectral Linear Iterative Clustering (CTF-SLIC) - An algorithm which is proposed in this thesis as an improvement of SLIC, combining it with the coarse-to-fine update strategy…

…2.3.2 Coarse-To-Fine Superpixel Segmentation (CTF) In contrast to SLIC, in which the algorithm operates on the entire full-size image, some algorithms employ a coarse-to-fine strategy. A coarse-to-fine strategy involves assigning superpixel…

…restricting label updates to the superpixel edges, the number of performed calculations can be heavily reduced, resulting in faster segmentation speeds. The Coarse-To-Fine superpixel segmentation algorithm (referred to as CTF in the sequel) is based…

…on the algorithms Coarse to Fine Monocular Segmentation (Yao et al. 2015) [24] and SEEDS (Van der Bergh et al. 2013) [22], with some modifications. The most noteworthy modifications are described below. This is…

.