Person Re-Identification with Improved Performance by Incorporating Focal Tversky Loss in AGW Baseline.

Person re-identification (re-ID) is one of the essential tasks for modern visual intelligent systems to identify a person from images or videos captured at different times, viewpoints, and spatial positions. In fact, it is easy to make an incorrect estimate for person re-ID in the presence of illumination change, low resolution, and pose differences. To provide a robust and accurate prediction, machine learning techniques are extensively used nowadays. However, learning-based approaches often face difficulties in data imbalance and distinguishing a person from others having strong appearance similarity. To improve the overall re-ID performance, false positives and false negatives should be part of the integral factors in the design of the loss function. In this work, we refine the well-known AGW baseline by incorporating a focal Tversky loss to address the data imbalance issue and facilitate the model to learn effectively from the hard examples. Experimental results show that the proposed re-ID method reaches rank-1 accuracy of 96.2% (with mAP: 94.5) and rank-1 accuracy of 93% (with mAP: 91.4) on Market1501 and DukeMTMC datasets, respectively, outperforming the state-of-the-art approaches.

Vision-Based Learning from Demonstration System for Robot Arms.

Robotic arms have been widely used in various industries and have the advantages of cost savings, high productivity, and efficiency. Although robotic arms are good at increasing efficiency in repetitive tasks, they still need to be re-programmed and optimized when new tasks are to be deployed, resulting in detrimental downtime and high cost. It is therefore the objective of this paper to present a learning from demonstration (LfD) robotic system to provide a more intuitive way for robots to efficiently perform tasks through learning from human demonstration on the basis of two major components: understanding through human demonstration and reproduction by robot arm. To understand human demonstration, we propose a vision-based spatial-temporal action detection method to detect human actions that focuses on meticulous hand movement in real time to establish an action base. An object trajectory inductive method is then proposed to obtain a key path for objects manipulated by the human through multiple demonstrations. In robot reproduction, we integrate the sequence of actions in the action base and the key path derived by the object trajectory inductive method for motion planning to reproduce the task demonstrated by the human user. Because of the capability of learning from demonstration, the robot can reproduce the tasks that the human demonstrated with the help of vision sensors in unseen contexts.

Cost Effective Mobile Mapping System for Color Point Cloud Reconstruction.

Survey-grade Lidar brands have commercialized Lidar-based mobile mapping systems (MMSs) for several years now. With this high-end equipment, the high-level accuracy quality of point clouds can be ensured, but unfortunately, their high cost has prevented practical implementation in autonomous driving from being affordable. As an attempt to solve this problem, we present a cost-effective MMS to generate an accurate 3D color point cloud for autonomous vehicles. Among the major processes for color point cloud reconstruction, we first synchronize the timestamps of each sensor. The calibration process between camera and Lidar is developed to obtain the translation and rotation matrices, based on which color attributes can be composed into the corresponding Lidar points. We also employ control points to adjust the point cloud for fine tuning the absolute position. To overcome the limitation of Global Navigation Satellite System/Inertial Measurement Unit (GNSS/IMU) positioning system, we utilize Normal Distribution Transform (NDT) localization to refine the trajectory to solve the multi-scan dispersion issue. Experimental results show that the color point cloud reconstructed by the proposed MMS has a position error in centimeter-level accuracy, meeting the requirement of high definition (HD) maps for autonomous driving usage.

Deep Learning-Based Real-Time Multiple-Person Action Recognition System.

Action recognition has gained great attention in automatic video analysis, greatly reducing the cost of human resources for smart surveillance. Most methods, however, focus on the detection of only one action event for a single person in a well-segmented video, rather than the recognition of multiple actions performed by more than one person at the same time for an untrimmed video. In this paper, we propose a deep learning-based multiple-person action recognition system for use in various real-time smart surveillance applications. By capturing a video stream of the scene, the proposed system can detect and track multiple people appearing in the scene and subsequently recognize their actions. Thanks to high resolution of the video frames, we establish a zoom-in function to obtain more satisfactory action recognition results when people in the scene become too far from the camera. To further improve the accuracy, recognition results from inflated 3D ConvNet (I3D) with multiple sliding windows are processed by a nonmaximum suppression (NMS) approach to obtain a more robust decision. Experimental results show that the proposed method can perform multiple-person action recognition in real time suitable for applications such as long-term care environments.

Iterative Pose Refinement for Object Pose Estimation Based on RGBD Data.

Accurate estimation of 3D object pose is highly desirable in a wide range of applications, such as robotics and augmented reality. Although significant advancement has been made for pose estimation, there is room for further improvement. Recent pose estimation systems utilize an iterative refinement process to revise the predicted pose to obtain a better final output. However, such refinement process only takes account of geometric features for pose revision during the iteration. Motivated by this approach, this paper designs a novel iterative refinement process that deals with both color and geometric features for object pose refinement. Experiments show that the proposed method is able to reach 94.74% and 93.2% in ADD(-S) metric with only 2 iterations, outperforming the state-of-the-art methods on the LINEMOD and YCB-Video datasets, respectively.

Distance measurement based on pixel variation of CCD images.

This paper presents a distance measurement method based on pixel number variation of CCD images by referencing to two arbitrarily designated points in the image frames. By establishing a relationship between the displacement of the camera movement along the photographing direction and the difference in pixel count between reference points in the images, the distance from an object can be calculated via the proposed method. To integrate the measuring functions into digital cameras, a circuit design implementing the proposed measuring system in selecting reference points, measuring distance, and displaying measurement results on CCD panel of the digital camera is proposed in this paper. In comparison to pattern recognition or image analysis methods, the proposed measuring approach is simple and straightforward for practical implementation into digital cameras. To validate the performance of the proposed method, measurement results using the proposed method and ultrasonic rangefinders are also presented in this paper.

Digital redesign of uncertain interval systems based on time-response resemblance via particle swarm optimization.

In this paper, a particle swarm optimization (PSO) based approach is proposed to derive an optimal digital controller for redesigned digital systems having an interval plant based on time-response resemblance of the closed-loop systems. Because of difficulties in obtaining time-response envelopes for interval systems, the design problem is formulated as an optimization problem of a cost function in terms of aggregated deviation between the step responses corresponding to extremal energies of the redesigned digital system and those of their continuous counterpart. A proposed evolutionary framework incorporating three PSOs is subsequently presented to minimize the cost function to derive an optimal set of parameters for the digital controller, so that step response sequences corresponding to the extremal sequence energy of the redesigned digital system suitably approximate those of their continuous counterpart under the perturbation of the uncertain plant parameters. Computer simulations have shown that redesigned digital systems incorporating the PSO-derived digital controllers have better system performance than those using conventional open-loop discretization methods.