Udaya Wijenayake

Dual pseudorandom array technique for error correction and hole filling of color structured-light three-dimensional scanning

Abstract. The accurate and dense real-time acquisition of three-dimensional (3-D) data using a low-cost structured light system remains an ongoing topic in the computer vision community, as it is difficult to achieve all these features simultaneously. Among several techniques, the pseudorandom array technique is widely used in real-time 3-D scene capturing, as it tends to concentrate the entire coding scheme into a single pattern. However, existing pseudorandom array decoding methods cannot decode a given symbol in real time when missing neighbors exist. As a solution, we propose a dual pseudorandom array encoding and decoding method and a hole-filling method, which can improve the reconstruction accuracy and time. We experimentally compared our method with several others to verify that our method captures 3-D scenes quickly and densely.

An Error Correcting 3D Scanning Technique Using Dual Pseudorandom Arrays

Designing a low cost structured light system which can acquire 3D data in real time with great accuracy is still an ongoing topic among computer vision community as it is hard to achieve all these features together. Among various structured light systems, pseudorandom array is the most suitable technique for real time 3D reconstruction as it tends to concentrate the entire coding scheme in a single pattern. But it has the difficulty of decoding the pattern when it loses few pattern symbols. In this paper, we address this problem by introducing an error correcting 3D scanning technique. This technique consists of a two way decoding method which can decode a dual pseudorandom array and a hole-filling algorithm. In the experimental results, we show that our 3D scanning technique largely improves the reconstruction compared to the conventional methods.

Real-Time External Respiratory Motion Measuring Technique Using an RGB-D Camera and Principal Component Analysis

Accurate tracking and modeling of internal and external respiratory motion in the thoracic and abdominal regions of a human body is a highly discussed topic in external beam radiotherapy treatment. Errors in target/normal tissue delineation and dose calculation and the increment of the healthy tissues being exposed to high radiation doses are some of the unsolicited problems caused due to inaccurate tracking of the respiratory motion. Many related works have been introduced for respiratory motion modeling, but a majority of them highly depend on radiography/fluoroscopy imaging, wearable markers or surgical node implanting techniques. We, in this article, propose a new respiratory motion tracking approach by exploiting the advantages of an RGB-D camera. First, we create a patient-specific respiratory motion model using principal component analysis (PCA) removing the spatial and temporal noise of the input depth data. Then, this model is utilized for real-time external respiratory motion measurement with high accuracy. Additionally, we introduce a marker-based depth frame registration technique to limit the measuring area into an anatomically consistent region that helps to handle the patient movements during the treatment. We achieved a 0.97 correlation comparing to a spirometer and 0.53 mm average error considering a laser line scanning result as the ground truth. As future work, we will use this accurate measurement of external respiratory motion to generate a correlated motion model that describes the movements of internal tumors.

Combination of Color and Binary Pattern Codification for an Error Correcting M-array Technique

Much research has been conducted so far to find a perfect structured light coding system. Among them, spatial neighborhood techniques which use a single pattern have become more popular as they can be used for dynamic scene capturing. But the difficulties of decoding the pattern when it loses few pattern symbols still remain as a problem. As a solution for this problem we introduce a new strategy which encodes two patterns into a single pattern image. In our particular experiment, we show that our decoding method can decode the pattern even it has some lost symbols.

PCA based analysis of external respiratory motion using an RGB-D camera

Human respiration induces considerable external and internal motion in the thoracic and abdominal regions. Tracking and modeling of this motion is an important task for accurate treatment planning and dose calculation during external beam radiotherapy. Inaccurate motion tracking can cause severe issues such as errors in target/normal tissue delineation and increment in the volume of healthy tissues exposed to high doses. Different methods have been introduced to model the respiratory motion, but most of them use wearable markers or surgical node implanting techniques, which are inconvenient to patients. In this paper, we experiment the feasibility of using an RGB-D camera along with Principal Component Analysis (PCA) to track and model the subject-specific external respiratory motion. Marker-based depth frame registration technique is also introduced to limit the measuring area into an anatomically consistent region during the treatment. We evaluate the accuracy of the proposed method using a Spirometer and a laser line scanner.

Stereo Vision-Based 3D Pose Estimation of Product Labels for Bin Picking

In the field of computer vision and robotics, bin picking is an important application area in which object pose estimation is necessary. Different approaches, such as 2D feature tracking and 3D surface reconstruction, have been introduced to estimate the object pose accurately. We propose a new approach where we can use both 2D image features and 3D surface information to identify the target object and estimate its pose accurately. First, we introduce a label detection technique using Maximally Stable Extremal Regions (MSERs) where the label detection results are used to identify the target objects separately. Then, the 2D image features on the detected label areas are utilized to generate 3D surface information. Finally, we calculate the 3D position and the orientation of the target objects using the information of the 3D surface.

Automatic detection and decoding of photogrammetric coded targets

Close-range Photogrammetry is widely used in many industries because of the cost effectiveness and efficiency of the technique. In this research, we introduce an automated coded target detection method which can be used to enhance the efficiency of the Photogrammetry.

Respiratory motion estimation using visual coded markers for radiotherapy

Respiratory motion is an important issue in radiotherapy as it can cause many problems such as damages to healthy tissues. Different methods have been introduced to estimate the respiratory motion but most of them need some electronic devices or expensive materials. In this research we proposed a respiratory motion estimation method by tracking inexpensive and easy to use visual coded markers using stereo vision.

Efficient extrinsic calibration of a laser range finder and camera using multiple edge registration

In this paper, we introduce an efficient way for extrinsic calibration between a LRF and a camera. By utilizing multiple edges of a known 3D structure, we derived a simple but robust calibration method. Through 3D reconstruction experiments, we verified the accuracy of the proposed method.

Head pose tracking using GPU based real-time 3D registration

The head pose tracking is one of the important criteria for improving the abilities of the human computer interactions and the human robot interactions. With the improvement of low cost consumer depth cameras lot of research attention attracted to the 3D based head pose estimation which is more accurate and robust to the environment conditions. In this paper we propose a head pose tracking method using a real-time 3D registration technique. In the experiment results we show that our method can handle larger pose variance accurately and robustly even without any kind of training or previous knowledge of the subject.