Denis Bose
Software Developer
Software Developer
Heidelberger Druckmaschinen
- Project: Print Inspection
- Implemented features and optimizations for an algorithm to detect defects on printed substrates of an offset printing press
- Implemented a method to detect smearing effects
- Integrated the algorithms into existing inspection software
- Skills: C++, OpenCV
Software Developer
Strabag ISS GmbH
- Project: Blind Detection
- Ported and extended an algorithm for detecting covered surveillance cameras to a current SDK
- Implemented new features for a video management system
- Skills: C++, Linux, Docker, C#
- Platforms: Windows
Software Architect
Carl Zeiss SMT GmbH
- Project: Particle Detection
- Developed a prototype to identify defects and particles on photomasks. Images were captured with a scanning electron microscope.
- Skills: C#, C++, OpenCV
- Tools: Git, Azure, Windows
Software Developer
Precitec GmbH & Co. KG
- Project: Image Stitching
- Implemented filters for control software for laser welding systems
- Analyzed an existing algorithm to create a large image from multiple smaller shots
- Defined optimization steps to improve the algorithm and reduce artifacts
- Skills: C++, Mercurial, KDevelop
Software Developer
Minebea Intec GmbH
- Project: CUPS Embedded
- Integrated CUPS print system software into the user interface of a weighing terminal running Yocto Linux
- Adapted the Yocto recipe for CUPS and related filters
- Developed and maintained software to generate labels (XML) and convert them to printable PDFs on the terminal
- Implemented an embedded PDF renderer that takes proprietary XML input and outputs a PDF for CUPS processing
- Skills: C++, Qt, Python
- Tools: Eclipse, Yocto
Software Developer
Concept Laser GmbH
- Project: 3D Print Monitoring
- Built QA software to visually monitor the powder bed of a 3D metal printer with a camera and document each step. Also applied image enhancement techniques.
- Skills: OpenCV, C++, Scrum, Jenkins, Windows
- Tools: Git, OPC UA
Software Developer
ASM Assembly Systems GmbH & Co. KG
- Project: Remote Stations Software
- Created a Windows service interface to read data from Siplace pick-and-place machines in the “Remote Stations” project. It retrieves installed software, version, and error data over LAN and sends it to another PC.
- Skills: C#, WCF, REST, Windows
Software Developer
Identt GmbH
- Project: ID-Card Reader
- Developed a prototype to read ID cards using a camera. The image is preprocessed so data fields can be read with OCR (Tesseract). Also built a tool to adapt the algorithm for different ID cards.
- Skills: OpenCV, Python, C++, C#
- Tools: Tesseract
Software Developer
AMAN digital.Media.Solutions
- Project: Trouble-Shooter
- Set up a development environment for parts of a complex distributed system to debug code efficiently when log analysis alone wasn’t enough. The app was in C#/Mono and PHP, hosted in Docker.
- Developed cloud services
- Database development
- Automated testing
- Skills: Docker, Git, Scrum, Python, SQL, C++, Eclipse, C#
Software Architect, Software Developer
Securiton GmbH (Permanent)
Project: Linux Embedded (Role: Software Architect, Oct 2016 - Mar 2017)
Ported automated monitoring algorithms to cameras from HikVision, Hanwha, and Axis. Each runs embedded Linux on different hardware. The app is identical across camera platforms and developed using Scrum.
Configured Eclipse for various targets
Wrote shell scripts to automate builds
Integrated TFS into the Linux environment
Designed a unified architecture for different camera vendors
Created a platform-independent library for ARM, Intel, Linux, and Windows
Set up an efficient desktop development and debugging environment
Streamed camera video to the desktop environment
Skills: POSIX, TFS, Eclipse, C++, Video
Platforms: ARM, Embedded Linux, HTTP
Project: CoreTester-GUI (Role: Software Developer, Oct 2016 - Nov 2016)
Developed a simple GUI for a test framework in C# using Windows Forms
Skills: Microsoft Visual Studio
Tools: Windows Forms
Software Developer
Mühlbauer AG (Permanent)
- Project: Photo Comparison
- Verified print quality of laser-etched photos. Challenges included varying scaling and brightness.
- Calibrated gray levels using a known color chart
- Located the photo using a correlator
- Determined exact scaling of the etched photo
- Scaled the photo accordingly
- Transformed the photo to another color space if needed (LBP, Canny, Sobel, HSV)
- Compared reference and etched photos
- Classified detected defects
- Skills: C++, OpenCV, Windows
Software Developer
Private
Project: Optical Character Recognition (Jan 2015 - Dec 2015)
Developed a size- and font-invariant OCR for normal characters using a support vector machine. The algorithm has a training and recognition phase.
Training: generate synthetic characters in various fonts
Training: scale characters to a uniform size
Training: compute gradients of characters
Training: train an SVM with computed gradients
Recognition: detect blobs in the image
Recognition: scale blobs to a uniform size
Recognition: compute gradients of blobs
Recognition: classify blobs as letters with the SVM
Skills: OpenCV, Support Vector Machine, Machine Learning
Project: OCR GUI (Aug 2015 - Dec 2015)
Built a WPF GUI to configure the OCR
Skills: WPF
Software Developer
Mühlbauer AG
- Project: GUI Design
- Designed a GUI with Microsoft Foundation Classes (MFC)
- Tools: MFC
Software Developer
Mühlbauer AG (Permanent)
- Project: ISO 1831
- Ensured that passport MRZ print quality meets ISO 1831. Character templates come from the OCR-B TrueType font.
- Checked contrast of the black character area
- Checked minimum stroke width of a character
- Checked maximum stroke width of a character
- Checked contrast of a character’s skeleton
- Checked for dirt in a character’s light area
- Checked for spots in a character’s dark area
- Checked correct alignment of a character
- Checked correct size of a character
- Skills: C++, OpenCV, Qt
Software Developer
Private
- Project: Android Game
- Developed a retro-style Asteroids game for Android
- Basic 2D rendering of moving objects
- Sound integration
- User interaction
- Tools: Android, Eclipse
Software Developer
Mühlbauer AG (Permanent)
- Project: Dotted OCR
- Improved an algorithm that reads dotted serial numbers (like in passports). It often misreads characters. Designed and implemented a new size- and rotation-invariant algorithm.
- Training and recognition have two phases:
- Training: segment dots with adaptive thresholding
- Training: normalize dots per character / remove scaling
- Training: each character has a set of 5–10 coordinates describing the dotted letter
- Recognition: segment dots with adaptive thresholding
- Recognition: cluster dots by morphological operations (dilate + erode) to group each letter’s dots
- Recognition: translate dots of a letter to the origin
- Recognition: normalize dots / remove scaling
- Recognition: find rotation to match recognition phase dots to training dots
- The new algorithm is scale- and rotation-invariant with zero misreads. It’s also configurable for performance or quality in poor images.
- Skills: OpenCV, C++
Debugger
Electrobit GmbH (Permanent)
- Project: ADTF
- Fused camera data for traffic sign recognition with navigation system data. The recognized sign is shown on a vehicle display. On test drives, if a wrong sign is detected, the recorded video is sent to the developer.
- Scene analysis with ADTF
- Skills: ADTF, C++, ADAS
Software Developer
Master’s Thesis
- Project: Object Recognition in Videos
- Implemented model-based algorithms to detect and locate objects in videos.
- Algorithm 1: built a 3D model
- Algorithm 1: calculated keypoints on the 3D model on the fly, including 3D coordinates
- Algorithm 1: at runtime, detected keypoints in the image
- Algorithm 1: matched image keypoints to model keypoints with RANSAC
- Algorithm 1: computed object position relative to the camera
- Algorithm 2: built a geometric line model of the object
- Algorithm 2: used Hough transform to detect lines in the image
- Algorithm 2: found a transform mapping model lines to image lines
- Algorithm 2: verified transformed lines against image lines
- Skills: OpenCV, Mathematica
Software Developer
Linearis3D GmbH
Project: Binary Coded Targets (Nov 2008 - Jul 2009)
Implemented an algorithm to detect binary coded circular markers for photogrammetry calibration.
Used least-squares to detect circles in the image
Checked low-error circles for binary coding by finding bright blobs at specific circle positions
Mapped blobs to bits and verified the code with a checksum
Project: Stereo System (Jan 2009 - Dec 2009)
Developed a video-based stereo system with a C# GUI
Detected markers in images from two calibrated cameras
Matched markers between images
Computed 3D coordinates via bundle adjustment
Live display of camera images in the GUI
Displayed 2D/3D coordinates in a table and as a 3D model with OpenGL
Skills: C#, Windows Forms, OpenGL
Project: Industrial Camera Control (Jan 2009 - Apr 2009)
Controlled and integrated industrial cameras into a software project
Integrated AVT cameras
Software Developer
Project Work / Paper
- Project: Sparse Bundle Adjustment Speedup Strategies
- Used Bundler to calibrate cameras without markers, speeding up the process via parallelization and splitting videos.
- Split the video into parts
- Each part contains images also in the other parts
- Bundler computes external camera parameters
- Transformed camera poses into one coordinate system using shared images
- Tools: Qt
Tester
David GmbH (Student)
- Project: Defining Software Tests
- Defined regression tests.
Bachelor’s Thesis
- Project: Stock Data via DAB
- Implemented the DAB-MOT protocol and a GUI in Java to visualize real-time stock data.
- Fetched real-time data via web service and uploaded it to an NDR server for DAB broadcasting.
- Skills: Java 1.5, Axis, Web Services, SOAP
I have extensive experience in software development with C++, C#, Java, and Matlab. My main focus so far has been on computer vision and image processing. I have designed and implemented several algorithms for challenging image processing tasks. I also have hands-on experience with machine learning and embedded Linux development. I learn quickly and solve algorithmic problems in a creative, empirical way. My solid knowledge lets me get up to speed in new areas fast.
Technische Universität Braunschweig
MSc Computer Science, focus on Robotics, Distributed Systems, Programming and Reactive Systems, specializing in Image Processing · Computer Science · Braunschweig, Germany
University of Osnabrück
BSc Mathematics/Computer Science, general, practical and theoretical computer science, foundations of mathematics · Mathematics/Computer Science · Osnabrück, Germany
ISTQB: Certified Tester
Technische Universität Braunschweig
Similar Freelancers
Discover other experts with similar qualifications and experience
