Technical Publications


Name of Technology	Automatic inspection technology for golf surface defects
Developer of Technology	Department of Electrical Engineering Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	The inspector will proceed the inspection over the uneven and non-flat surface of a golf ball to detect several types of surface defects – for example, dents, burs and scratch which damages the surface. For a given 11 types of various defect, the detection rate is about 97% and speed of the inspection is about 23 balls per minute.


Name of Technology	AOI Detection Technology for Appearance Defects of Memory Module Products
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	Proceed AI model training with deep learning network framework, and develop AI technology for “Defective Image Classification”. The accuracy rate of defect classification is more than 99%, which effectively solves the problem of high misjudgment ratio of traditional optical inspection equipment and the need for a large number of manual re-examinations.

Third Prize

First Prize

Best Innovation Award


Name of Technology	Chamfering Tool Diagnosis System
Developer of Technology	Department of Electrical Engineering Chung-Wen Hung, Associate Professor
Summary of Technology	It is currently used in the diagnosis of chamfering tools. In the future, by reading the processing signals of different machines, and retraining the model that meets the application from the collected data, then detect the abnormality. Achieve the same anomaly detection effect, improve the overall efficiency and yield of the production line.

Ching-Lung Chang

Professor, Department of Computer Science and Information Engineering
Vice Dean, College of Engineering
Director, Graduate School of Engineering Science and Technology

Laboratory

Embedded Network System Laboratory

Chuan-Yu Chang

Distinguished Professor, Department of Computer Science and Information Engineering
Director, IRIS Research Center
Deputy CEO, Service Systems Technology Center, ITRI

Laboratory

Medical Image Processing Laboratory

Research Topics

Machine learning/Deep learning/ Neural networks

Experience

Chief Technology Officer, Service Systems Technology Center, ITRI (2019/08-now)

Director, IRIS Research Center, YunTech (2018/01-now)

President, Taiwan Association for Web Intelligence Consortium (2017/10-now)

Director, Image Processing and Pattern Recognition Society (2014/01-now)

Distinguished Professor, Department of Computer Science and Information Engineering, YunTech (2013/08-now)

Dean, Office of Research and Development, YunTech (2011/08-2019/07)

Director, Incubation Center for Academia-Industry Collaboration and Intellectual Property, YunTech (2011/08-2019/07)

Chair, Department of Computer Science and Information Engineering, YunTech (2009/08-2011/07)

Ching-Ju Chen

Project-Appointed Assistant Professor, Bachelor Program in Interdisciplinary Studies

Ting-Lung Lin

Assistant Professor, Department of Architecture and Interior Design

Bang-Ning Hwang

Associate Professor and Chairperson, Department of Business Administration

Hui-Ling Huang

Associate Professor, Department of Applied Foreign Languages

Yueh Hsiu Giffen Cheng

Professor and Chairperson, Department of Creative Design

Yung-Kuan Tseng

Professor, Department of Cultural Heritage Conservation

Deng-Chuan Cai

Distinguished Professor, Department of Industrial Design
Vice Dean, College of Design
Dean, Graduate School of Design-Doctoral Program

Chung-Chian Hsu

Distinguished Professor, Department of Information Management

Laboratory

Intelligent Database System Laboratory

Ching-Lung Su

Associate Professor, Department of Electronic Engineering
Section Director, Library-System Information Group

Laboratory

System on Chip Laboratory

Shr-Chang Shia

Distinguished Professor, Department of Electronic Engineering

Laboratory

Advanced Circuit and Chip Design Laboratory

Ming-Hwa Sheu

Distinguished Professor, Department of Electronic Engineering

Laboratory

Integrated Circuit System Integration Laboratory

Chian C. Ho

Associate Professor, Department of Electrical Engineering
Section Director, YunTech Office of Research and Development- Employment and Alumni Liaison Group

Laboratory

Embedded System on Chip Laboratory

Chung-Wen Hung

Associate Professor, Department of Electrical Engineering

Laboratory

Advanced Purpose Integrated Circuits and Systems Design Laboratory

EHsien-Huang P. Wu

Distinguished Professor, Department of Electrical Engineering
Director, Intelligent and Monitoring Research Center
Director, Power Electronics and Sustainable Energy Technology R & D Center

Laboratory

Automated Optical Inspection Laboratory

Edward T.-H. Chu

Associate Professor, Department of Computer Science and Information Engineering

Laboratory

Embedded Operation System (Envision on System) Laboratory

Wen-Fong Wang

Associate Professor, Department of Computer Science and Information Engineering
Section Director, YunTech Office of Research and Development-Research Service Group

Laboratory

E-Health Informatics & System Laboratory

Shih-Yu Chen

Associate Professor, Department of Computer Science and Information Engineering

Laboratory

Hyperspectral Signal and Image Processing Laboratory

Chien-Chou Lin

Associate Professor, Department of Computer Science and Information Engineering

Laboratory

Embedded Intelligent System Laboratory

Arthur, Chang

Project-Appointed Assistant Professor, Bachelor Program in Interdisciplinary Studies

Chang Hsuan-Ting

Professor, Department of Electrical Engineering

Related Links

Photonics and Information Lab


Name of Technology	Facial-Heart Rate and Stroke Detection System
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	By recognizing facial muscle change, such as crooked eyes, squinted eyes and other functions to defect the risk of facial stroke. In order to observe the micro-vibration of the head due to the contraction of the heart, develop a non-contact facial heart rate detection technology.
Video	<span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start"></span>


Name of Technology	Drunk Driving Sensing System based on Heartbeat Micro – Quiver Sensing Technology
Developer of Technology	Department of Computer Science and Information Engineering Wen-Fong Wang, Associate Professor
Summary of Technology	The system can measure the heartbeat vibrations by setting it on the steering wheel or other places that are in direct contact with the body, or it can be applied to personal 3C items such as smart bracelets. This technology can be installed on the steering wheel or seat in car, with the alarm system and cloud database, it can immediately detect whether the driver is drinking or not, and prevent this kind of behavior or notify affected members to the scene to deal with it.


Name of Technology	Using Hyperspectral Technology Develop Agriculture 4.0 with Coffee Bean Quality Management as an Example
Developer of Technology	Department of Computer Science and Information Engineering Shih-Yu Chen, Associate Professor
Summary of Technology	In the past, coffee merchants used manpower and eyes to pick defective coffee beans, for example: moths. Now it can be identified through intelligent image recognition, also can reduce manpower, costs, and wrong judgments.
Video


Name of Technology	Thyroid Image Diagnosis Technology
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	This technology assists doctors during clinical diagnosis to reduce the difficulty in diagnosis, decrease the inspection time and improve the diagnosis rate. The contributions are as follows. (1) Realization of automatic thyroid gland segmentation in various medical images, (2) Segmentation and classification of thyroid nodules in ultrasound images, (3) Volume estimation of the thyroid gland in ultrasound images, (4) Segmentation of thyroid tumors and (5) automatic detection of the Graves disease in the thyroid gland.


Name of Technology	Infant Crying Translator
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	Before learning how to express the emotions or physiological/psychological requirements with language, infants usually express how they feel to parents through crying. We have collected more than 2,000,000 infant cries. Machine learning and the big data analysis techniques are used to automatically adapt the crying model for each infant. The cloud system is built to collect and analyze the cries. Infant Crying Translator recognizes four types of infant crying, including being hungry, in pain, wet-diaper and sleepy. Experimental results have revealed good performance of the proposed system, the recognition accuracy is up to 92%, 85.4%, 83.8%, and 77% for infants born within 2 weeks, 1 month, 2 months, and 4 months, respectively.
Video


Name of Technology	Smart Baby Monitor
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	Obstruction of the nose and mouth by vomit or quilts is the main cause of sudden infant death. Different baby sleeping postures will increase the difficulty of face detection. Therefore, a deep learning neural network is adopted to detect baby faces and a novel algorithm is proposed to detect vomiting of infants. The proposed method can detect baby’s face effectively under the complex background and various illumination conditions. The system can help parents monitor the baby whether his/her mouth and nose is covered with vomit or quilt. The system can real-time notification to parents, to improve infant safety and reduce the burden on parents.


Name of Technology	Tire Bubble Defect Detection System
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	This study proposes a method of combining convolutional neural networks and Faster R-CNN classifiers to automatically detect bubble defects in tire speckle images. The proposed method reduces misjudgments caused by human errors and achieves high consistency in the detection quality. The experimental results show that in addition to completely detecting the bubble defects, the proposed method also reduces the false positive rate of non-bubble defects.


Name of Technology	Automatic Defect Detection Of Wood Board
Developer of Technology	Department of Electrical Engineering Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	The system is based on computer stereo vision, Which uses the triangulation method to detect the defect by a line laser and an industrial camera. Putting the wood board on a mobile platform for movement and a line of laser is projected onto the surface of the board, the depth image of the board surface can then be reconstructed to 1mm accuracy. The depth image can be used for defect detection by simple image processing, and the defects detected are used for performing subsequent glue filling operation or cutting operation.


Name of Technology	Shoe Surface 3D Reconstruction and Path Planning for Glue Spraying
Developer of Technology	Department of Electrical Engineering Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	The technology combines computer vision with the laser triangulation technology to build a 3D scanning system, which generates automatic glue spraying path information according to the computations and analysis performed by the algorithm. The information is transmitted to a six-axial robotic arm, shoe-surface glue-spraying system to automate and standardize the glue spraying process. The technology automates the glue-spraying path modelling to increase the efficiency by 80% and enables the possibility of small-scale production of multiple products.


Name of Technology	3D Visual Identification and Orientation Estimation of Large Objects
Developer of Technology	Department of Electrical Engineering Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	Using an inexpensive 3D depth camera, the system can capture the depth information of the scene. The depth information is used to construct the surface geometry of the 3D object and perform feature extraction. The feature extraction can be matched to the existing 3D CAD model to identify the classification and calculate the position and orientation angle of the object. These three types of information can finally be transferred back to the 6-axis robot for automatic pick-and-place action.


Name of Technology	AI Monitoring System for Safe Behavior of Entrance/Exit on Production Line
Developer of Technology	Director, IRIS Research Center Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	Technology content: Important issue of industrial safety accident prevention, to avoid the employees who do not follow the safety procedures and seek the convenience to cause industrial safety incidents. Therefore, a set of AI Monitoring System for Safe Behavior of Entrances/Exits on Production Line was developed, using pedestrian tracking and behavior recognition technology to timely monitor whether employees follow safety standards and improve the safety of employees’ operations.


Name of Technology	Extreme 3D modeling and recognition technology
Developer of Technology	IRIS research center Director Department of Computer Science and Information Engineering Chuan-Yu Chang, Distinguished Professor
Summary of Technology	The technology comprises of advanced optical zoom, face tracking as well as 3D face biometric technology and can be applied to 3D face recognition, cosmetic surgery and orthodontics.
Video


Name of Technology	Solar Power Plant Abnormal Timely Detection and Diagnosis System
Developer of Technology	Department of Information Management Chung – Chian Hsu, Distinguished Professor
Summary of Technology	The data collected in series from the solar panels are transmitted through the network, stored in a cloud database, and analyzed by machine learning technology to detect and diagnose equipment abnormalities. After the abnormalities of solar panels is detected and diagnosed immediately, the situation will be texted through mobile phone to notify maintenance engineers arrived at the scene instantly and repair the power losses to a minimum.


Name of Technology	Vehicle Surround View Display System with Free Viewing Function(AVM)
Developer of Technology	Department of Electronic Engineering Ching-Lung Su, Associate Professor
Summary of Technology	4 optical wide-angle fisheye lenses are used to provide a panoramic view of the vehicle, showing high-resolution images of the entire vehicle from above and at any angle of view. When parking and reversing, obstacles around the vehicle can be inspected without blind spot, improving safety and protection.


Name of Technology	Cultural Relics and Artwork Intelligent Recognition Design
Developer of Technology	Department of Cultural Heritage Conservation Yung – Kuan Tseng, Professor
Summary of Technology	Use intelligent recognition technology to develop unique identification tools for antiquities or artworks created by specific artists or painters. Boosting cultural development is based on a sound cultural and artwork market, and the security of transactions depends on authenticity – fake recognition, preserve truly valuable antiquities.
Video


Name of Technology	UAV intelligent recognition technology
Developer of Technology	Bachelor Program in Interdisciplinary Studies Ching-Ju Chen, Project-appointed Assistant Professor
Summary of Technology	The technology uses the UAV as the carrier to perform intelligent recognition on images. For example, to perform rapid and accurate AI positioning for the thousands of oyster frames on the sea and to find the locations and quantity with the data from the GPS.
Video


Name of Technology	An intelligent recognition system for a UAV flight carrier with high-altitude moving target
Developer of Technology	Bachelor Program in Interdisciplinary Studies Ching-Ju Chen, Project-appointed Assistant Professor
Summary of Technology	The technology integrates the UAV’s terrain independent mobility with the stable flying ability and the YOLO v3 target recognition technology to provide applications for the agricultural industry, environmental monitoring, emergency search and rescue, scientific research, resource exploration, homeland security and ecological monitoring.


Name of Technology	Mowing robot
Developer of Technology	Department of Computer Science and Information Engineering Ching-Lung Chang, Professor
Summary of Technology	Embedded platform, combined with RTK-GPS, 9-axis sensor and the Kalman filter, PID control, ultrasonic sensing, image recognition and other technologies. Create a mowing robot with precise autonomous mowing and obstacle avoidance capability. Relevant technologies can be transferred to achieve automatic mowing and save manpower. It can be applied to lawn finishing in campus, parks, homes, etc.
Video


Name of Technology	IOT Application that Implement Automatic Photography with Sound Recognition
Developer of Technology	Department of Electronic Engineering Ming-Hwa Sheu, Distinguished Professor
Summary of Technology	Detecting and recognizing the sound of clinking glasses during “Cheers! “, it will automatically activate the camera and take a picture from the sound source, these photos will be uploaded to the voice control device named ” IOT Application that Implement Automatic Photography with Sound Recognition”, the simple and handy design allows various parties such as wedding venues and bars to receive the most natural and vivid gathering photos.


Name of Technology	Driver Monitoring System
Developer of Technology	Department of Computer Science and Information Engineering Associate Professor, Ching-Lung Su
Summary of Technology	The driver monitoring system (DMS) enables to automatically monitor the physical and psychological conditions of the driver while driving. The monitored contents vary from the differences of the system, normally including fatigue-driving, smoking, using mobile phones, talking on the phone, paying attention to the road, physiological emergencies, etc.


Name of Technology	Intelligent Trash Can Featuring Patrolling and Surveillance
Developer of Technology	Department of Electrical Engineering Chian-C. Ho, Associate Professor
Summary of Technology	Traditional trash cans are usually installed in the out-of-the-way dead corners so that people are often littering the trash anywhere because of urgency or inconvenience. This work “Intelligent Trash Can Featuring Patrolling and Surveillance” can collect the trach proactively from people so as to improve the effectiveness and efficiency of cleaning service. On the other hand, traditional security surveillance cameras are usually installed on the ceiling corners or crossways so that there are still many dead spaces and security loopholes to be monitored. This work can patrol and inspect proactively so as to improve the effectiveness and efficiency of surveillance service.


Name of Technology	Design and Development of International Disease Classification Code ICD-10 Coding System
Developer of Technology	Bachelor Program in Interdisciplinary Studies Jung-Sheng Chang, Project-appointed Assistant Professor
Summary of Technology	Use big data analysis technology to help encode the doctor’s diagnosis of medical information into ICD-10 to improve the accuracy and speed of coding and reduce costs. Develop an out-patient clinic intelligent medical pre-diagnosis system that allows patients to answer questions in advance through mobile phones or multi-media interactive machines before the consultation, assist doctors to conduct more in-depth consultations during limited visits time to improve medical quality.


Name of Technology	Automatic Drug Identification System
Developer of Technology	Department of Electrical Engineering, Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	Combining computer vision with Raspberry pi, and using relays to control the two cameras to take pictures of the front and back of the tablet, it can automatically identify the name and the number of pills, and reduce the mistakes made by health care worker when taking pills and supplement, because the pills are too similar, avoid causing patients to ingest the wrong medicine.


Name of Technology	Indoors Gait Positioning System (Indoors GPS)
Developer of Technology	Department of Computer Science and Information Engineering Wen-Fong Wang, Associate Professor
Summary of Technology	The indoors GPS applies the cell-phone built-in sensors (accelerometer, magnetometer, gyroscope, etc.) with dead reckoning algorithms to locate users, who stay in a complex indoors separated space. This novel idea has been developed in a cell-phone user APP. This app has the functions of service initialization, indoor navigation, check-in, and more.


Name of Technology	Community Health Care and Fitness Tracking and Service System
Developer of Technology	Department of Computer Science and Information Engineering Wen-Fong Wang, Associate Professor
Summary of Technology	The system is capable of monitoring multiple heath data, evaluate and record the health status of the elderly and provide the user with all of the relevant information regarding the elderly through a mobile app.


Name of Technology	Gait Analysis System with Intelligent Image Recognition
Developer of Technology	Department of Computer Science and Information Engineering Wen-Fong Wang, Associate Professor
Summary of Technology	This technology can accurately record and analyse the human gait and be applied to neurology to nervous system inspection.
Video


Name of Technology	Copper Foil Defect Detection
Developer of Technology	Department of Electrical Engineering, Hsien-Huang P. Wu , Distinguished Professor
Summary of Technology	Traditional AOI is able to inspect 23 types of defects. Using feature descriptors, classification is difficult and time-consuming. To solve this problem we introduce AI deep learning to perform direct classification. Preliminary research result shows that the accuracy in defect classification is 93%, which continues to improve with the availability of more data. This technology can replace existing inspection equipment (11 stages) and reduce the cost of the original equipment by 50%.


Name of Technology	A Platform for the Construction of Shoe Last Point Cloud Models
Developer of Technology	Department of Computer Science and Information Engineering Chien-Chou Lin, Associate Professor
Summary of Technology	The technology uses a mobile platform with a line laser scanner to obtain the shoe last’s point cloud data from all angles. The point cloud data are merged using point cloud merging algorithms to obtain the complete shoe last’s point cloud model.


Name of Technology	Image Pre-Processing for Tire Defect Detection — Image Calibration Based on the Fusion of Depth and Color Images
Developer of Technology	Department of Computer Science and Information Engineering Chien-Chou Lin, Associate Professor
Summary of Technology	The technology integrates a laser displacement meter and a high-speed color camera. The tire surface’s depth variation is captured by a laser displacement meter and combined with the tire surface’s visual images captured by a colour camera. The two types of information are rectified and fused to obtain the texture characteristics from the images and the physical attributes from the depth images to improve the accuracy of tire defect detection.


Name of Technology	Point Cloud Object Recognition and Tracking Algorithm
Developer of Technology	Department of Computer Science and Information Engineering Chien-Chou Lin, Associate Professor
Summary of Technology	Laser light’s visual ability are stronger than image, it won’t affected by light illumination, and won’t misjudged shadow as an object, this technology can identify the type of object , such as cars or pedestrians, it can also track 360 degree moving objects.
Video


Name of Technology	Using Hyperspectral Technology to Estimate Vase Life of Anthurium Cut
Developer of Technology	Department of Computer Science and Information Engineering Shih-Yu Chen, Associate Professor
Summary of Technology	The use of hyperspectral information with multi-band and non-interaction characteristics combined with machine learning methods to estimate the vase life of anthurium and accurately predict the aging time of anthurium to increase market value.
Video

2019's Technical Publications

2019/05/28 Smart Identification Technology Achievement Conference

2019/06/28 Formosa Plastics Group.4th Innovation Conference

2019/10/18 IRIS Technology Conference

2019/11/20 Automotive ADAS Product Presentation and Press conference

2019/11/30 Intelligent UAV Image Recognition Technology and Application

2019/12/04 Press Conference on AOI Technology