Improved Feature Matching Research Articles

Extraction of Visual Landmarks Using Improved Feature Matching Technique for Stereo Vision Applications

Stereo vision is becoming more and more common in three-dimensional visualization, autonomous vehicle navigation, path finding, object recognition, and other computer vision applications. In this paper, we present an approach for extracting visual landmarks from images acquired by a stereoscopic camera. The scale invariant feature transform (SIFT) technique with self-organizing map is used to detect and recognize visual landmarks. Our methodology is based on winner calculation technique, and the main idea is to keep in the database only distinctive features or landmarks in order to minimize detection time. We will demonstrate that this methodology is more efficient than ordinary SIFT algorithm or speeded up robust features (SURF) matching technique. Improved feature group matching with computation time better than SIFT and the SURF has been observed in the experiments with a variety of stereo image pairs.

Depth estimation of features in video frames with improved feature matching technique using Kinect sensor

Estimating depth has long been a major issue in the field of computer vision and robotics. The Kinect sensor's active sensing strategy provides high-frame-rate depth maps and can recognize user gestures and human pose. This paper presents a technique to estimate the depth of features extracted from video frames, along with an improved feature-matching method. In this paper, we used the Kinect camera developed by Microsoft, which captured color and depth images for further processing. Feature detection and selection is an important task for robot navigation. Many feature-matching techniques have been proposed earlier, and this paper proposes an improved feature matching between successive video frames with the use of neural network methodology in order to reduce the computation time of feature matching. The features extracted are invariant to image scale and rotation, and different experiments were conducted to evaluate the performance of feature matching between successive video frames. The extracted features are assigned distance based on the Kinect technology that can be used by the robot in order to determine the path of navigation, along with obstacle detection applications.

The MOPED framework: Object recognition and pose estimation for manipulation

We present MOPED, a framework for Multiple Object Pose Estimation and Detection that seamlessly integrates single-image and multi-image object recognition and pose estimation in one optimized, robust, and scalable framework. We address two main challenges in computer vision for robotics: robust performance in complex scenes, and low latency for real-time operation. We achieve robust performance with Iterative Clustering Estimation (ICE), a novel algorithm that iteratively combines feature clustering with robust pose estimation. Feature clustering quickly partitions the scene and produces object hypotheses. The hypotheses are used to further refine the feature clusters, and the two steps iterate until convergence. ICE is easy to parallelize, and easily integrates single- and multi-camera object recognition and pose estimation. We also introduce a novel object hypothesis scoring function based on M-estimator theory, and a novel pose clustering algorithm that robustly handles recognition outliers. We achieve scalability and low latency with an improved feature matching algorithm for large databases, a GPU/CPU hybrid architecture that exploits parallelism at all levels, and an optimized resource scheduler. We provide extensive experimental results demonstrating state-of-the-art performance in terms of recognition, scalability, and latency in real-world robotic applications.