Salient Region Detection in outdoor scene using Imagery and 3-D Data

People

Publication

Gunhee Kim, Daniel Huber, and Martial Hebert
Segmentation of Salient Regions in Outdoor Scenes using Imagery and 3-D Data
IEEE Workshop on Application of Computer Vision (WACV 2008), Colorado, USA, January 7-9, 2008. (Oral)
[Paper(PDF)] [Presentation(PPT)]

Description

This project develops an algorithm for detecting and segmenting out the salient regions of a scene using both 3-D laser scan and imagery information. Fig.1 shows our problem statement. Given an image and its extrinsically calibrated 3-D scan data, our algorithm delineates top k-most salient regions in a bottom-up manner without any high-level priors, models, or learning.

The detection of salient clusters is formulated into finding the set of clusters that best fit some pre-defined families of probability density functions in 3-D scan data. Here we use two pdfs, which are Gaussian and Uniform. As a fundamental tool for this approach, we use the Robust Information-theoretic Clustering (RIC) method. After clustering, we calculate saliency value of each cluster by summing four different saliency features, which capture local, regional, and global properties of 3-D data and RGB color information.

Fig.2 shows some examples using the dataset taken in parking lots around CMU.

Fig.1. Problem Statement. (a) Input: an image and its corresponding 3-D scan data. (b) Information-theoretic optimal clustering of 3-D scan data. (c) Output: segmentation of the top 4 most salient regions.

Fig.2. More examples of the detection of salient regions. The first column shows the input images. The second column represents the results of RIC clustering. The third column is the top four most salient regions in the 3-D data. The fourth column shows their segmentation in the image.

Funding

Robotics Consortium sponsored by the U.S Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD19-01-2-0012.