A5062081080- Advanced Neural Network Applications
- Domain Adaptation and Few-Shot Learning
- Video Surveillance and Tracking Methods
- Gaze Tracking and Assistive Technology
- Human-Automation Interaction and Safety
- Tissue Engineering and Regenerative Medicine
- Older Adults Driving Studies
- Social Robot Interaction and HRI
- Remote Sensing and Land Use
- Simulation and Modeling Applications
- Urban Transport and Accessibility
- Image Enhancement Techniques
- Multimodal Machine Learning Applications
- Cellular Mechanics and Interactions
- Muscle Physiology and Disorders
- Cardiomyopathy and Myosin Studies
- Robotics and Sensor-Based Localization
- Context-Aware Activity Recognition Systems
- Human Pose and Action Recognition
- Teleoperation and Haptic Systems
- 3D Printing in Biomedical Research
- Tactile and Sensory Interactions
- Machine Learning and Data Classification
- Remote-Sensing Image Classification
- Image and Video Stabilization
Clemson University
2020-2023
This paper proposes a real-time multi-scale semantic segmentation network (MsNet). MsNet is combination of our novel fusion with matching attention model (MFMA) as the decoding and searched by asymptotic neural architecture search (ANAS) or MobileNetV3 encoding network. The MFMA not only extracts low level spatial features from inputs but also decodes contextual extracted ANAS. Specifically, considering advantages disadvantages addition concatenation fusion, we design (MF) that balances...
Recently significant progress has been made in 3D detection. However, it is still challenging to detect small contour objects under complex scenes. This paper proposes a novel Attention-based Multi-phase Multi-task Fusion (AMMF) that uses point-level, RoI-level, and multi-task fusions complement the disadvantages of LiDAR camera, solve this challenge. First, at feature extraction phase, AMMF Low High-level with Matching Attention (LHF-MA) efficient FPN (eFPN) perform point-level fusion for...
Assistive navigation for blind or visually impaired (BVI) individuals is of significance to extend their mobility and safety in traveling, enhancing employment opportunities fostering personal fulfillment. Conventional research mainly based on robotic approaches through localization, mapping, path planning frameworks. They require heavy manual annotation semantic information maps its alignment with sensor mapping. Inspired by the fact that we human beings naturally rely language instruction...
Traditional manual wheelchairs have a fixed seat with no movement or angle adjustment, which can seriously affect the user's comfort and greatly limit user experience. However, electric wheelchair relies on strong intelligence automatic features; it not only realize multidegree freedom adjustment of human body but also has rich powerful man–machine control interface, facilitates improves This study upgraded Permobil C400-powered multisensor data fusion technology to enrich its terrain...
Assistive robots can be found in hospitals and rehabilitation clinics, where they help patients maintain a positive disposition. Our proposed robotic mobility solution combines state of the art hardware software to provide safer, more independent, productive lifestyle for people with some most severe disabilities. New includes, retractable roof, manipulator arm, hard backpack, number sensors that collect environmental data processors generate 3D maps hands-free human-machine interface.The...
This research proposes a novel semi-supervised learning framework FourStr (Four-Stream formed by two two-stream models) that focuses on the improvement of fusion and labeling efficiency for 3D multi-sensor detector. adopts single-stage detector named adaptive network (AFNet) as backbone trains it through semi-supervision (SSL) strategy Stereo Fusion. Note AFNet SSL Fusion can benefit each other. On one hand, Four-stream composed AFNets naturally provides rich inputs large models While other...
Current learning-based 3-D object detection accuracy is heavily impacted by the annotation quality. It still a challenge to expect an overall high for all classes under different scenarios given dataset sparsity. To mitigate this challenge, article proposes novel method called semi-supervised learning and progressive distillation (SPD), which uses (SSL) knowledge improve label efficiency. The SPD two big backbones hand unlabeled/labeled input data augmented periodic IO augmentation (PA)....
In vitro cultures to be used in various analytical investigations of cardiomyocyte (CM) growth and function for enhancing insight into physiological pathological mechanisms should closely express vivo morphology. The aim the studies is explore how use microfabrication physical-cue-addition techniques establish a neonatal rat CM culture model that expresses an end-to-end connected rod shape with vivo-like intercalated discs (ICDs). Freshly isolated CMs were cultured on microgrooved...
The feature pyramid, which is a vital component of the convolutional neural networks, plays significant role in several perception tasks, including object detection for autonomous driving. However, how to better fuse multi-level and multi-sensor pyramids still challenge, especially detection. This paper presents FocusTR (Focusing on valuable features by multiple Transformers), simple yet effective architecture, pyramid single-stream 2D detector two-stream 3D detector. Specifically,...
Using our lab built two-photon excitation fluorescence (TPEF) and second harmonic generation (SHG) hybrid confocal imaging system, we observed, for the first time, dynamic sarcomeric addition process in a rat cardiomyocyte cell culture system; this system expressed vivo-like myofibril structure mimicked mechanical overload experienced heart muscle tissue. Micro-grooved topographic patterned substrates com- bined with electrical stimulation are used to achieve structure. After cardiomyocytes...
Individuals with visual impairments encounter many obstacles passenger vehicles. This study aimed to increase the understanding of challenges specifically related vehicles including ingress, in-vehicle considerations, comfort, and acceptance ridesharing transportation options for individuals who are visually impaired. Ten participants impaired, an average age 57.5 years, completed a semi-structured interview. The interview took place over Zoom or phone focused on their vehicle preferences...