skip to main content

Search for: All records

Creators/Authors contains: "Yin, Z."

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. The state-of-the-art of fully-supervised methods for temporal action localization from untrimmed videos has achieved impressive results. Yet, it remains unsatisfactory for the weakly-supervised temporal action localization, where only video-level action labels are given without the timestamp annotation on when the actions occur. The main reason comes from that, the weakly-supervised networks only focus on the highly discriminative frames, but there are some ambiguous frames in both background and action classes. The ambiguous frames in background class are very similar to the real actions, which may be treated as target actions and result in false positives. On the other hand, themore »ambiguous frames in action class which possibly contain action instances, are prone to be false negatives by the weakly-supervised networks and result in a coarse localization. To solve these problems, we introduce a novel weakly-supervised Action Completeness Modeling with Back- ground Aware Networks (ACM-BANets). Our Background Aware Network (BANet) contains a weight-sharing two-branch architecture, with an action guided Background aware Temporal Attention Module (B-TAM) and an asymmetrical training strategy, to suppress both highly discriminative and ambiguous background frames to remove the false positives. Our action completeness modeling contains multiple BANets, and the BANets are forced to discover different but complementary action instances to completely localize the action instances in both highly discriminative and ambiguous action frames. In the 𝑖-th iteration, the 𝑖-th BANet discovers the discriminative features, which are then erased from the feature map. The partially-erased feature map is fed into the (i+1)-th BANet of the next iteration to force this BANet to discover discriminative features different from the 𝑖-th BANet. Evaluated on two challenging untrimmed video datasets, THUMOS14 and ActivityNet1.3, our approach outperforms all the current weakly-supervised methods for temporal action localization.« less
  2. In a human-centered intelligent manufacturing system, every element is to assist the operator in achieving the optimal operational performance. The primary task of developing such a human-centered system is to accurately understand human behavior. In this paper, we propose a fog computing framework for assembly operation recognition, which brings computing power to the data source, to achieve real-time recognition. The operator’s activity is captured using visual cameras. Instead of directly training a deep learning model from scratch, transfer learning is applied to transfer the learning abilities to our application. A worker assembly operation dataset is established, which at present containsmore »10 sequential operations in an assembly task of installing a desktop CNC machine. The developed model is evaluated on this dataset and achieves a recognition accuracy of 95% in the testing experiments.« less
  3. Training and on-site assistance is critical to help workers master required skills, improve worker productivity, and guarantee the product quality. Traditional training methods lack worker-centered considerations that are particularly in need when workers are facing ever changing demands. In this study, we propose a worker-centered training & assistant system for intelligent manufacturing, which is featured with self-awareness and active-guidance. Multi-modal sensing techniques are applied to perceive each individual worker and a deep learning approach is developed to understand the worker’s behavior and intention. Moreover, an object detection algorithm is implemented to identify the parts/tools the worker is interacting with. Thenmore »the worker’s current state is inferred and used for quantifying and assessing the worker performance, from which the worker’s potential guidance demands are analyzed. Furthermore, onsite guidance with multi-modal augmented reality is provided actively and continuously during the operational process. Two case studies are used to demonstrate the feasibility and great potential of our proposed approach and system for applying to the manufacturing industry for frontline workers.« less
  4. In today's competitive production era, the ability to identify and track important objects in a near real-time manner is greatly desired among manufacturers who are moving towards the streamline production. Manually keeping track of every object in a complex manufacturing plant is infeasible; therefore, an automatic system of that functionality is greatly in need. This study was motivated to develop a Mask Region-based Convolutional Neural Network (Mask RCNN) model to semantically segment objects and important zones in manufacturing plants. The Mask RCNN was trained through transfer learning that used a neural network (NN) pre-trained with the MS-COCO dataset as themore »starting point and further fine-tuned that NN using a limited number of annotated images. Then the Mask RCNN model was modified to have consistent detection results from videos, which was realized through the use of a two-staged detection threshold and the analysis of the temporal coherence information of detected objects. The function of object tracking was added to the system for identifying the misplacement of objects. The effectiveness and efficiency of the proposed system were demonstrated by analyzing a sample of video footages.« less
  5. Production innovations are occurring faster than ever. Manufacturing workers thus need to frequently learn new methods and skills. In fast changing, largely uncertain production systems, manufacturers with the ability to comprehend workers' behavior and assess their operation performance in near real-time will achieve better performance than peers. Action recognition can serve this purpose. Despite that human action recognition has been an active field of study in machine learning, limited work has been done for recognizing worker actions in performing manufacturing tasks that involve complex, intricate operations. Using data captured by one sensor or a single type of sensor to recognizemore »those actions lacks reliability. The limitation can be surpassed by sensor fusion at data, feature, and decision levels. This paper presents a study that developed a multimodal sensor system and used sensor fusion methods to enhance the reliability of action recognition. One step in assembling a Bukito 3D printer, which composed of a sequence of 7 actions, was used to illustrate and assess the proposed method. Two wearable sensors namely Myo-armband captured both Inertial Measurement Unit (IMU) and electromyography (EMG) signals of assembly workers. Microsoft Kinect, a vision based sensor, simultaneously tracked predefined skeleton joints of them. The collected IMU, EMG, and skeleton data were respectively used to train five individual Convolutional Neural Network (CNN) models. Then, various fusion methods were implemented to integrate the prediction results of independent models to yield the final prediction. Reasons for achieving better performance using sensor fusion were identified from this study.« less