Embedded AI Deep Learning Technology for ADAS/Niche Self-Driving Applications
| Summary | Principal Investigator:Professor Jiun-In Guo Summary To assist Taiwan’s automotive industry to break through the barrier established by the existing ADAS patent portfolio in the world, we propose the idea of developing embedded AI deep learning technology for ADAS/niche self-driving applications so that Taiwan’s industry possess the chance to be part of the eco-system of ADAS/Self-driving products in the world. Our research focuses on three areas: (1) Automatic data labeling tool for deep learning applications and labeled ADAS datasets; (2) Embedded deep learning object detection/behavior analysis algorithm and model development; (3) Real-time deep learning computing platform development. Applications of our research include autonomous driving functions like lane keeping, automatic emergency braking (AEB), and automatic lane changing. In addition, we are also developing related technology required in certain niche self-driving applications, e.g., non-tracking AGV highly demanded in the Industry 4.0, to satisfy the needs of local industry. Keywords Embedded deep learning technology, fast data labeling, object detection and behavior recognition, ADAS, autonomous vehicle, industry 4.0 Innovations We have developed a world first fast labeling tool, ezLabel (shown in Fig. 1), to speed up 10x labeling efficiency as compared to the existing manual video labeling tools, which won two prizes in 2018 AUDI Innovation Award Taiwan. We have established a dataset with 11M+ of samples for deep learning object detection and behavior analysis for ADAS/Self-driving applications (as shown in Fig. 2). Out of the collected 11M+ samples, we have opened and shared 96K samples of them, most of which are related to vehicles, pedestrians, and cyclists under various weather conditions. We have developed an embedded deep learning model that can detect vehicles as far as 200m, which outperforms the Yolo v2 model (max. 50m) and owns over 10% higher in mAP as well. We have developed an embedded SSD lite model (SSD 512×512) that is suitable for using the TI TDA2X chipset for real-time object recognition. At 30fps, we obtained 72% mAP@Pascal VoC (as shown in Fig. 3). We have developed a deep learning technology based on 3D convolution neural network to predict whether rear vehicles will overtake in the next three seconds with a 95% or more accuracy (as shown in Fig. 4). We have developed a Taiwan first smart self-driving wheelchair for indoor mobility in places like airports and hospitals (as shown in Fig. 5). We have developed a world first bit accurate dynamic fixed-point deep learning model training and inferencing tool for CNN hardware accelerators. We have developed a hybrid fixed-point/binary CNN model training flow for CNN hardware accelerators (as shown in Fig. 6). Benefits Have won two prizes in 2018 AUDI Innovation Award Taiwan. Have won an outstanding prize in AISlander 2018 with ezLabel and embedded AI technology. Have won a Technology Breakthrough Award in MOST 2018 Future Tech. Have won a Bronze award in 2018 MXIC design contest. Have conducted 18 cases of Industrial collaboration projects with total grant about NT$15,980K. |
||
|---|---|---|---|
| Keyword | TES TEST TEST TEST TEST | ||
| Download | test.pptx | ||
| Research Project | |||
| Research Team | |||
More like this
Provide the latest information of AI research centers and applied industries
-
An integrated manufacturing platform, the law of sciencetechnology,industrial ecosystem - smart productionintelligent precision manufacturing with digital decision, AI modeling, big data governancekernel technologies
This project aims to develops a completely manufacturing flexible decision framework by applying big data analytics and AI techniques that integrates data from different decision-making units such as APC, APS, capacity planning, and inventory management planning to increase the decision quality and resilient capability in uncertain risks based on the perspective of factory operation.
-
An integrated manufacturing platform, the law of sciencetechnology,industrial ecosystem - smart productionintelligent precision manufacturing with digital decision, AI modeling, big data governancekernel technologies
A self-improving thermal error compensation model is established by federated learning to improve the machine tool performance under different ambient temperatures. Each machine can share and update the latest model parameter to improve the initial model from different ambient temperature variation trails, but still keep the private ambient information of each machine.
-
Defect detection and re-polishing system for highly reflective polished and ground parts.
Develop manufacturing technology and value for difficult-to-machine superalloy materials, fully automated machining technology, improve efficiency and machining quality stability, and combine automation and intelligence for unmanned autonomous operation of machining and inspection.
-
Integration of an ICU Visualization Dashboard (i-Dashboard) as a Platform to Facilitate Multidisciplinary Rounds
Multidisciplinary rounds (MDRs) are scheduled, patient-focused communication mechanisms among multidisciplinary providers in the intensive care unit (ICU). The surgical ICU team of National Cheng Kung University Hospital has developed and integrated i-Dashboard as a platform to facilitate MDRs. i-Dashboard is a custom-developed visualization dashboard that supports (1) key information retrieval and reorganization, (2) time-series data, and (3) display on large touchscreens during MDRs. The i-Dashboard increases the efficiency in data gathering and enhances communication accuracy and information exchange in MDRs.
-
Super-fast Convergence for Radiance Fields Reconstruction
The NeRF-based technique describes a super-fast convergence approach to reconstructing the per-scene radiance field from a set of images that capture the scene with known poses.
-
Machine Learning to Predict In-Hospital Cardiac Arrest in Patients Admitted from the Emergency Department with COVID-19 and Suspected Pneumonia
By using the machine learning algorithms, this study developed a risk stratification model for predicting the occurrence of in-hospital cardiac arrest (IHCA) events in patients admitted from the emergency department with COVID-19 and pneumonia. The results showed that the model's performance is better than by using the National Early Warning Score (NEWS).
-
Cardiovascular Health Guardian – Novel Pulse Wave Velocity and Personal Blood Pressure Estimation System for Smart Watch
Our team develops an accurate PWV estimation algorithm that uses wrist PPG and ECG signals from wearable devices. A missing-feature imputation and ambiguous-feature resolution technique is developed and the availability of wrist PPG morphological features is raised from 60% to 99.1%. A weighted pulse decomposition approach is adopted and 5 component waves can be acquired to examine more detailed properties. The PWV is then estimated by XGBoost algorithm with the hierarchical regression model.
-
![HeaortaNet (Automatic Pericardium/Aorta Segmentation AI Model [HeaortaNet])](https://www.innoaitw.org/website/uploads_product/website_1/P0000100000383_4_694.jpg?_96854 "HeaortaNet (Automatic Pericardium/Aorta Segmentation AI Model [HeaortaNet])")
HeaortaNet (Automatic Pericardium/Aorta Segmentation AI Model [HeaortaNet])
The Pericardium/Aorta Segmentation and Cardiovascular Risk Prediction AI Total Solution Model, HeaortaNet, is a deep learning model based on UNet and attention gate, and had been trained by >70,000 axial images with verified annotations of the pericardium and aorta. It shortens the time for data processing from 60 minutes, by manual segmentation of both pericardium and aorta, to 0.4 seconds. The segmentation accuracy is 94.8% for the pericardium, and 91.6% for the aorta. The applicability of HeaortaNet had been demonstrated by analyzing the non-contrast chest CT scans (>5,000 cases) deposited in the mega-image bank of National Health Insurance Databank.
-
Embedding multimodal machine intelligence in the digital life of AI technology
This project collaborates with the international team to collect a very large-scale Chinese emotional corpus. In terms of technology, the fairness of speech emotion recognition is also discussed to solve social issues that may be encountered regarding the usability of emotion recognition. Among them, it is found that the database annotations are all labeled with the unfair perspective of men and women, which leads to biases in the trained model. In order to solve this problem, there have been preliminary achievements in the technological development of fairness, and will be submitted in the near future.
