Image quality software engineer with expertise in image processing & computer vision and software/firmware engineering in diverse industries like Consumer electronics, Automotive electronics and slot machine companies which range in size from small company to large multinational corporations.
Key areas of expertise and skills include:
- Image processing
- Computer vision
- Data visualization
- Object Orientation Programming
- Embedded System
- IP Camera
- Video Decoder
- Camera 3A algorithm optimization
- Project planning/management and evaluation
- System Design
- Management and Training
My passion for work influences colleagues to become positive to solve problems.
Image Quality Software Engineer
Taipei Special Municipality，TW
● Camera Product design, Camera Image quality tuning and analyze (PDAF, CAF and TOF tuning)
● Experience in C/C++ coding as well as software system design (Qualcomm S865, camx, chi-cdk).
● Camera 3A (Auto focus) tuning within top 3 ranks of DxOMark Mobile (project: VIVO X50 Pro+, published date: Oct. 2020).
● Experience in Python coding as well as shell scripting.
* 如對Vivo工作有興趣，可幫忙內推 [email protected]
● Establish software structure, user interface and laser measurement library
● Developing multi-thread GUI software check for Wires, rods, bars, tubes and profiles in factory.
● Developing Data visualization software by QT creator(c/c++, Object Orientation Programming) on the Linux and windows platform
● Developing calibration algorithm and communication protocol for dimensional measurements in steel mills, reading out measurement sensors through TCP/IP or RS485 and evaluating their data.
● Collaborating and training our Service Engineer in Foreign Parent company
● Familiar with Nokia Eagles Project camera 3A (mainly focused on AF) algorithm optimization tuning (Qualcomm S630/660)
● Experience in Nokia Dragon Project camera driver porting
● Implement of Nokia Panda Project camera shutter feature on the MTK 6771 platform
● Training junior firmware engineer to bring up MCU Controller Area Network on the after-sales service project(CAN Bus) (Microchip & Freescale chip)
● Training software team to develop 3D vehicle model on the Sketch up pro platform (3D module build software).
● Enhance Japan customer(JVC KENWOOD) Vehicle electronics video protocol drivers (HDMI protocol).
● Development of 3D Around View Monitor System (AVM) (Hyundai H1，Nissan Tiida & New Sentra)
● Eagle view system consists of one ECU and four cameras synthesizes, enabling the driver to steer into a parking space with ease
● Familiar with DSP chip including TI DM648, Fujitsu MB86R11/12 series，Avisonic AVS7110
● Work experience involved developing the CVR(Car Video Recoder) System used Sunplus 6330 chipset
● Developing PC player relatived to the CVR product.
行車紀錄器(DVR)開發整合衛星導航功能(GPS)以及語音測速警示(eDog)功能，行車紀錄檔的PC多媒體軟體開發，以及Hunting Camera專案開發，產品主要使用凌陽Sunplus 5330/6330晶片組。
● Duties involved developing monitoring and recording applications system (Car Video Recoding System) on the arm-linux platform.
● 16ch-DVR/NVR Project mainly used TI 8148 chips and encode/decode multimedia network video packets (H.264 / MJPEG), along with both FFmpeg open-source code and LIVE555 open-source code.
16ch-DVR/NVR ffmpeg影音編解碼程式撰寫，網路連線程式撰寫，PC播放程式撰寫(開發工具-Visual Studio)，IP Camera韌體程式撰寫(晶片-TI DM368)，車用監控系統開發(晶片-TI 8148)。
● Develop the software and firmware on SLOT machine.
● Duties involved researching and developing the SAS(Slot Accounting System) for overseas customers through the arm architecture.
● Transmitting encrypted financial data from customers’ stores to their headquarters by SMS, UART, and via TCP/IP.
● The job entailed working in Digital Video recorder development, and researching IP-TV.
Advanced skill in image processing, computer vision,
C/C++, Matlab, Object Orientation Programming, OpenGL, RS485, Network
Advanced skill in SoC/Embedded Linux development, from kernel and up. NVR, Video server, IP Camera, Video Decoder, H.264, JPEG, MPEG4, RTSP/RTP, Network protocol includes ONVIF, PSIA, and Embedded System (e.g. Microchip, Freescale)
Texture plays an important role in human perception and medical image area. In this year, I implement a novel texture descriptor, local parameter histograms (LPHs), based on Gaussian Markov random fields (GMRFs). Significant Image segmentation result of 2D image are achieved and will be applied on the three-dimensional texture.
Design of Intelligent IP Camera Middleware
The evolution of the surveillance system has been a migration from the analog CCTV (Closed Circuit Television) system to the integration system of digital IP camera. The design of intelligent IP camera middleware is proposed and implemented based on a reliable architecture of surveillance software system. The main contributions of the thesis as follows. The image searching algorithm for surveillance system is purposed. There are two subsystems including face recognition subsystem and face/object comparison subsystem in the algorithm. In the algorithm, four major functions comprise skin color recognition, image noise removal, connected component labeling, and face or object comparison. The image searching algorithm for surveillance system can be applied to the design of surveillance software platform in the future.
安全監控系統以及感應式的門禁系統的應用，可以防範保護居家環境，並達到防止宵小等入侵之目的。感應式的門禁系統只能單純感應門禁磁卡，無法辨別宵小身分，提供更近一步的管制動作，致使許多社區、公司行號、居家大樓住戶除了使用安全監控系統以及感應式門禁系統之外，仍需另外雇請保全人員監視出入口安全，導致成本相對提高。因此，本研究的目的，提出一個「具智慧型影像辨識功能之演算法設計與實作」，改善傳統感應式門禁系統的不足，本研究包含2 個子系統，人臉辨識子系統與人臉或物件比對子系統，其中包含膚色辨識 、雜訊去除、連通物件演算法與人臉或物件比對四大功能。智慧型影像辨識功能之演算法，可支援未來IP Camera 應用系統之設計與安全監控系統之開發應用的彈性。
Universal Middleware Bridge System (UMBS) for IP cam networking is designed and proposed. The UMBS provides the related mechanisms for system manual setting, automatic configuration, and management to improve the whole procedures of setting and installation. The UMBS is composed of four main functional modules, Live Video (LV), Playback Video (PV), Intelligent Scheduler (iScheduler), and System Configuration (SC). The structure of the robust UMBS supports the adaptability and flexibility for design and development of IP cam application system. The intelligent surveillance appliances system integrated with the robust UMBS can be realized efficiently in the future.
Keywords: Universal Middleware Bridge System (UMBS), IP Camera, ONVIF, Middleware, Platform.
The evolution of the surveillance system has been a migration from the analog CCTV (Closed Circuit Television) system to the integration system of digital IP camera. The design of intelligent IP camera middleware is proposed and implemented based on a reliable architecture of surveillance software system. The main contributions of the thesis as follows. The image searching algorithm for surveillance system is purposed. There are two subsystems including face recognition subsystem and face/object comparison subsystem in the algorithm. In the algorithm, four major functions comprise skin color recognition, image noise removal, connected component labeling, and face or object comparison. The image searching algorithm for surveillance system can be applied to the design of surveillance software platform in the future. Keywords: Intelligent IP Camera, Surveillance, Image Recognition.
Three-Dimensional Gaussian Markov Random Field based improved texture descriptor for image segmentation