More Like this

1. Robust Activity Recognition for Adaptive Worker-Robot Interaction using Transfer Learning

   Shahnavaz, Farid; Tavassoli, Riley; Akhavian, Reza (January 2023, 2023 ASCE International Conference on Computing in Civil Engineering)

   Turkan, Yelda; Louis, Joseph; Leite, Fernanda; Ergan, Semiha (Ed.)

   Human activity recognition (HAR) using machine learning has shown tremendous promise in detecting construction workers’ activities. HAR has many applications in human-robot interaction research to enable robots’ understanding of human counterparts’ activities. However, many existing HAR approaches lack robustness, generalizability, and adaptability. This paper proposes a transfer learning methodology for activity recognition of construction workers that requires orders of magnitude less data and compute time for comparable or better classification accuracy. The developed algorithm transfers features from a model pre-trained by the original authors and fine-tunes them for the downstream task of activity recognition in construction. The model was pre-trained on Kinetics-400, a large-scale video-based human activity recognition dataset with 400 distinct classes. The model was fine-tuned and tested using videos captured from manual material handling (MMH) activities found on YouTube. Results indicate that the fine-tuned model can recognize distinct MMH tasks in a robust and adaptive manner which is crucial for the widespread deployment of collaborative robots in construction. 

   more » « less
2. Transferable Two-Stream Convolutional Neural Network for Human Action Recognition

   Xiong, Q. (January 2020, 48th North American Manufacturing Research Conference)

   Human-Robot Collaboration (HRC), which envisions a workspace in which human and robot can dynamically collaborate, has been identified as a key element in smart manufacturing. Human action recognition plays a key role in the realization of HRC as it helps identify current human action and provides the basis for future action prediction and robot planning. Despite recent development of Deep Learning (DL) that has demonstrated great potential in advancing human action recognition, one of the key issues remains as how to effectively leverage the temporal information of human motion to improve the performance of action recognition. Furthermore, large volume of training data is often difficult to obtain due to manufacturing constraints, which poses challenge for the optimization of DL models. This paper presents an integrated method based on optical flow and convolutional neural network (CNN)-based transfer learning to tackle these two issues. First, optical flow images, which encode the temporal information of human motion, are extracted and serve as the input to a two-stream CNN structure for simultaneous parsing of spatial-temporal information of human motion. Then, transfer learning is investigated to transfer the feature extraction capability of a pretrained CNN to manufacturing scenarios. Evaluation using engine block assembly confirmed the effectiveness of the developed method. 

   more » « less
3. Transferable two-stream convolutional neural network for human action recognition

   https://doi.org/10.1016/j/jmsy.2020.04.007  

   Xiong, Q. (January 2020, Journal of manufacturing systems)

   Human-Robot Collaboration (HRC), which enables a workspace where human and robot can dynamically and safely collaborate for improved operational efficiency, has been identified as a key element in smart manu­facturing. Human action recognition plays a key role in the realization ofHRC, as it helps identify current human action and provides the basis for future action prediction and robot planning. While Deep Learning (DL) has demonstrated great potential in advancing human action recognition, effectively leveraging the temporal in­formation of human motions to improve the accuracy and robustness of action recognition has remained as a challenge. Furthermore, it is often difficult to obtain a large volume of data for DL network training and opti­mization, due to operational constraints in a realistic manufacturing setting. This paper presents an integrated method to address these two challenges, based on the optical flow and convolutional neural network (CNN)­based transfer learning. Specifically, optical flow images, which encode the temporal information of human motion, are extracted and serve as the input to a two-stream CNN structure for simultaneous parsing of spatial­-temporal information of human motion. Subsequently, transfer learning is investigated to transfer the feature extraction capability of a pre-trained CNN to manufacturing scenarios. Evaluation using engine block assembly confirmed the effectiveness of the developed method. 

   more » « less
4. Multi-Modal Recognition of Worker Activity for Human-Centered Intelligent Manufacturing

   https://doi.org/10.1016/j.engappai.2020.103868  

   Tao, W.; Leu, M.; Yin, Z. (October 2020, Engineering applications of artificial intelligence)

   This study aims at sensing and understanding the worker’s activity in a human-centered intelligent manufacturing system. We propose a novel multi-modal approach for worker activity recognition by leveraging information from different sensors and in different modalities. Specifically, a smart armband and a visual camera are applied to capture Inertial Measurement Unit (IMU) signals and videos, respectively. For the IMU signals, we design two novel feature transform mechanisms, in both frequency and spatial domains, to assemble the captured IMU signals as images, which allow using convolutional neural networks to learn the most discriminative features. Along with the above two modalities, we propose two other modalities for the video data, i.e., at the video frame and video clip levels. Each of the four modalities returns a probability distribution on activity prediction. Then, these probability distributions are fused to output the worker activity classification result. A worker activity dataset is established, which at present contains 6 common activities in assembly tasks, i.e., grab a tool/part, hammer a nail, use a power-screwdriver, rest arms, turn a screwdriver, and use a wrench. The developed multi-modal approach is evaluated on this dataset and achieves recognition accuracies as high as 97% and 100% in the leave-one-out and half-half experiments, respectively. 

   more » « less
5. Leveraging Affect Transfer Learning for Behavior Prediction in an Intelligent Tutoring System

   https://doi.org/10.1109/FG52635.2021.9667001  

   Ruiz, Nataniel; Yu, Hao; Allessio, Danielle A.; Jalal, Mona; Joshi, Ajjen; Murray, Thomas; Magee, John J.; Whitehill, Jacob R.; Ablavsky, Vitaly; Arroyo, Ivon; et al (December 2021, 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021))

   In this work, we propose a video-based transfer learning approach for predicting problem outcomes of students working with an intelligent tutoring system (ITS). By analyzing a student's face and gestures, our method predicts the outcome of a student answering a problem in an ITS from a video feed. Our work is motivated by the reasoning that the ability to predict such outcomes enables tutoring systems to adjust interventions, such as hints and encouragement, and to ultimately yield improved student learning. We collected a large labeled dataset of student interactions with an intelligent online math tutor consisting of 68 sessions, where 54 individual students solved 2,749 problems. We will release this dataset publicly upon publication of this paper. It will be available at https://www.cs.bu.edu/faculty/betke/research/learning/. Working with this dataset, our transfer-learning challenge was to design a representation in the source domain of pictures obtained “in the wild” for the task of facial expression analysis, and transferring this learned representation to the task of human behavior prediction in the domain of webcam videos of students in a classroom environment. We developed a novel facial affect representation and a user-personalized training scheme that unlocks the potential of this representation. We designed several variants of a recurrent neural network that models the temporal structure of video sequences of students solving math problems. Our final model, named ATL-BP for Affect Transfer Learning for Behavior Prediction, achieves a relative increase in mean F -score of 50 % over the state-of-the-art method on this new dataset. 

   more » « less