Hello ;-)
Computer Vision · Machine Learning · Software Engineering
I'm Chadly Marouane, a software engineer based in Munich with a background in computer vision and machine learning. I earned my doctorate at Ludwig Maximilian University of Munich (LMU), where my research focused on visual positioning and human movement analysis. I currently work at TAWNY.
Research
My research focused on how cameras can help us understand where we are and how we move. I started with indoor positioning, using images captured by smartphones to recognize places and estimate a person's location.
Later, I explored how changes in visual details from one video frame to the next could reveal patterns of human movement. Inspired by approaches from sensor-based activity recognition, I investigated camera-based methods for detecting steps and recognizing activities.
Topics: Computer vision · Indoor positioning · Visual odometry · Human activity recognition
Selected research
- Indoor Positioning Using Smartphone CameraIEEE IPIN, 2011 · Using smartphone images to estimate indoor location and orientation.
- Visual Positioning Systems – An Extension to MoVIPSIEEE IPIN, 2014 · Further development of camera-based indoor positioning.
- Step and Activity Detection Based on the Orientation and Scale Attributes of the SURF AlgorithmIEEE IPIN, 2016 · Detecting steps and activities from changes in visual features over time.
- Visual Odometry Using Motion Vectors from Visual Feature PointsIEEE IPIN, 2016 · Estimating pedestrian motion using visual features.
- Visuelle Verfahren für ortsbezogene DiensteDoctoral dissertation, LMU Munich, 2017.
Apps
- GlucosePhone — Track your blood sugar values on iPhone & iPad
- GlucoseWatch — Track your blood sugar values on Apple Watch
- GlucoseDesk — Track your blood sugar values on macOS
- TranslateChadly — Private Offline AI Translation on iPhone & iPad
- Rower Tracker — Analyze your rowing workouts on iPhone & Apple Watch
Publications & Speaking
Complete list, including the selected works above.
- DeepFlow: Detecting Optimal User Experience From Physiological Data Using Deep Neural NetworksAug 16, 2019
- DeepFlow: Detecting Optimal User Experience From Physiological Data Using Deep Neural Networks (Extended Abstract)May 16, 2019
- An Open, Labeled Dataset for Analysis and Assessment of Human MotionAugust 24, 2018
- Automated Recognition and Difficulty Assessment of Boulder RoutesFebruary 17, 2018
- Methods for location, activity and context recognition based on OpenCVDecember 1, 2017
- Digitalization in Schools-Organization, Collaboration and CommunicationSeptember 15, 2017
- Visuelle Verfahren für ortsbezogene DiensteJuly 3, 2017
- Sensx: About sensing and assessment of complex human motionMay 17, 2017
- Visual odometry using motion vectors from visual feature pointsOctober 4, 2016
- Step and activity detection based on the orientation and scale attributes of the SURF algorithmOctober 4, 2016
- Visual Odometry for Pedestrians Based on Orientation Attributes of SURFSeptember 21, 2016
- Mobile Authentisierung im UnternehmensalltagFebruary 9, 2016
- SURFLogo - Mobile Tagging with App IconsNovember 1, 2015
- KeyPocket-Improving Security and Usability for Provider Independent Login Architectures with Mobile DevicesOctober 26, 2015
- Classification of Vehicle Types in Car Parks using Computer Vision TechniquesJuly 22, 2015
- Enabling Pedometers on Basis of Visual Feature Point ConversionApril 24, 2015
- Trends und Chancen beim mobilen EinkaufenApril 10, 2015
- Vis-a-Vis: Offline-Capable Management of Virtual Trust Structures Based on Real-Life Interactions2014
- Visual Positioning Systems - An Extension to MoVIPSOctober 27, 2014
- SURFtogether: Towards Context Proximity Detection Using Visual FeaturesOctober 07, 2014
- Free-hand pointing for identification and interaction with distant objectsOctober 28, 2013
- Vis-a-vis verification: Social network identity management through real world interactions2013
- Indoor positioning using smartphone cameraSeptember 21, 2011
Patents
- System for gesture-based adjustment of seat mounted in vehicle by user, has control unit that controls setting of vehicle seat associated with recognized gesture and gesture areaDE102012216181A1 - 2012
- System i.e. park assistance system, for controlling movement of motor car from outside by user, has control unit arranged according to detected gesture for controlling recognized gesture associated with movement of motor carDE102012216174A1 - 2012
- Method for determining location of object e.g. shop by user of vehicle, involves determining direction vector and object position data with respect to predetermined reference point and absolute position in geographic coordinate systemDE102012208733A1 - 2012
- System for controlling air-conditioner and infotainment device of i.e. car by user, has control unit for receiving signal indicating detected gesture from recognition unit and controlling recognized function of air conditioner and deviceDE102012216184A1 - 2012
- Method for operating massage device, has determining whether position of input unit corresponds to position in area assigned to operating element, and activating operating element when position of input unit corresponds to assigned positionDE102012211164A1 - 2012