www.acad.jobs : academic jobs worldwide – and the best jobs in industry
                
     
Position: Master thesis „Geometric cue learning from stereo depth data”
Institution: Austrian Institute of Technology GmbH
Location: Wien, Austria
Duties: Based on our recent research results one of your tasks is to extend our algorithmic pipeline towards using dense stereo depth data; Moreover, you will learn to investigate the possibilities and limits of depth databased detection of learned geometric cues in a bottom-up manner; You will learn to use graph-based representations to reconstruct the shapes of multiple, partially overlapping (and occluding) objects in real-time
Requirements: Ongoing master’s studies in the field of computer science, software engineering, information and computer engineering, mathematics, robotics or similar; Knowledge of fundamental Computer Vision and/or Machine Learning concepts; Experience in Python, preferably also in PyTorch; Interest in applied aspects of modern machine learning and robot vision; Good knowledge of verbal and written English
   
Text: We are Austria’s largest research and technology organisation and an international player in applied research for innovative infrastructure solutions. This makes us a powerful development partner for industry and a top employer in the scientific community. Our Center for Vision, Automation & Control in Vienna invites applications for a: Master thesis „Geometric cue learning from stereo depth data” Our team’s expertise expertise is sensor and algorithm development for automated vehicles and machines that perceive the environment in 3D. Robotic systems need abilities to recognize objects and to navigate in unknown environments. Geometry is a strong universal cue for the environment and its rigid objects, as it remains unchanged across different object appearances and viewpoints. Based on our recent research results one of your tasks is to extend our algorithmic pipeline towards using dense stereo depth data. Moreover, you will learn to investigate the possibilities and limits of depth databased detection of learned geometric cues in a bottom-up manner. You will learn to use graph-based representations to reconstruct the shapes of multiple, partially overlapping (and occluding) objects in real-time. You will elaborate a novel depth data pre-processing, data learning and inference framework. You will understand how to validate this framework on a set of increasingly complex practical scenarios and finalize it as a proof of concept demonstrator system in Python/PyTorch. You will learn how to evaluate and document test runs in pre-defined scenarios and publish the results at a robotics/computer vision conference or journal. Your qualifications as an Ingenious Partner* : Ongoing master’s studies in the field of computer science, software engineering, information and computer engineering, mathematics, robotics or similar Knowledge of fundamental Computer Vision and/or Machine Learning concepts Experience in Python, preferably also in PyTorch Interest in applied aspects of modern machine learning and robot vision Good knowledge of verbal and written English What to expect EUR 779,25 gross per month for 20 hours/week based on the collective agreement. There will be additional company benefits. You will be part of our international Young AIT network. As a research institution, we are familiar with the supervision and execution of master theses and we are looking forward to supporting you accordingly. At AIT, the promotion of women is important to us - that's why we are especially looking forward to applications from female students! Please submit your application documents including your CV, motivational letter and certificates online. Tomorrow Today - with You? Apply now! Apply online now Back to job listing Print × Close Interested? Drag your CV here or upload it to create a new profile. Upload your CV ;
Please click here, if the job didn't load correctly.
Your browser does not support iframes. Please click <a href="https://www.acad.jobs/job.php?t_id=J000341885&redirect" target="_parent" style="color:#7A7A7A">here</a>, if the job didn't load correctly.