Computer Vision Master's in Germany: 2026 Programme Guide

Computer vision in Germany is usually studied through a broader master's degree rather than a programme literally named "Computer Vision."

Depending on your goal, the right route may be:

• Visual Computing
• Computer Science with vision and graphics electives
• Machine Learning with vision projects
• Robotics with perception and sensor fusion
• Data and Computer Science
• Medical image computing

The university with the most visible vision lab is not automatically the best programme for you. First check admission eligibility, then prove that the current curriculum can support a coherent vision profile.

This guide was reviewed on 8 June 2026.

Short Answer

Before shortlisting universities, choose the technical profile you want to build:

Then reject any programme that fails one of these gates:

1. You meet its formal prior-degree and subject-credit requirements.
2. You can complete enough relevant modules in your usable language.
3. The modules are offered frequently enough to form a study plan.
4. You can access suitable projects, seminars, and thesis supervision.
5. Tuition and total cost are affordable without speculative employment.

Computer Vision Is Not One Discipline

A technically complete profile draws from several foundations.

Mathematics

• Linear algebra
• Multivariable calculus
• Probability and statistics
• Optimisation
• Numerical methods
• Geometry

Computer Science

• Programming
• Algorithms and data structures
• Software engineering
• Computer architecture and systems
• Databases or data management
• Parallel or distributed computing where relevant

Vision and Machine Learning

• Image formation and signal processing
• Classical feature and estimation methods
• Machine learning and deep learning
• Detection, segmentation, tracking, and recognition
• Multiple-view geometry and 3D reconstruction
• Evaluation, uncertainty, robustness, and domain shift

Systems and Domain Knowledge

• Cameras, optics, calibration, and sensors
• GPU or accelerator deployment
• Robotics and control
• Medical, automotive, manufacturing, remote-sensing, or other application knowledge
• Privacy, safety, security, and regulatory constraints

A programme consisting mainly of general AI modules may not provide enough geometry, imaging, systems, or domain depth for your target role.

Current Programme Examples

These examples show different programme architectures. They are not a ranking.

Saarland University: Visual Computing MSc

Saarland offers a dedicated, English-language Visual Computing MSc. Its current description combines image analysis and pattern recognition with image generation, geometry, image synthesis, mathematics, physics, informatics, and mechatronics.

Current admission controls include:

• A bachelor's or equivalent qualification in Visual Computing or a related field, particularly computer science, mathematics, physics, electrical engineering, or mechatronics, or equivalent academic achievement
• Evidence of particular interest through a dossier or qualified reference
• Advanced English at C1
• No acceptance of a medium-of-instruction letter as English proof
• Winter and summer entry

The programme page also identifies participation from Saarland informatics, mathematics, systems engineering, Fraunhofer biomedical engineering, and the Max Planck Institute for Informatics. Treat this as evidence of the programme's academic structure, not a guarantee of a particular supervisor or research position.

Official Saarland Visual Computing MSc page

Heidelberg University: Data and Computer Science MSc

Heidelberg's current degree is Data and Computer Science, not the older generic Computer Science title used in many lists. It is a consecutive, 120 ECTS programme that can be completed entirely in English.

The 2026 module handbook includes Computer Vision and describes subject areas including machine learning, visual computing, and data analysis. Admission is restricted by requirements and qualifications, including core computer-science preparation and English evidence under the current regulations.

This is a broad computing degree with a possible vision profile. Applicants must verify:

• The current admission regulation
• Which visual-computing modules are available in the intended semesters
• Whether enough electives and the thesis can form a coherent specialisation
• Baden-Wuerttemberg tuition and exemption status

Heidelberg's programme page lists the semester contribution separately. Many non-EU/EEA students in Baden-Wuerttemberg also pay the state tuition of EUR 1,500 per semester unless exempt.

Official Heidelberg Data and Computer Science MSc page

Official Heidelberg application information

University of Freiburg: Computer Science MSc

Freiburg offers a broad 120 ECTS Computer Science MSc with an open curriculum, an Artificial Intelligence specialisation, and a Cyber-Physical Systems specialisation.

The AI specialisation can include artificial intelligence, machine learning, deep learning, robotics, computer vision, computer graphics, and neuroscience. Its current module handbook includes an English Computer Vision module.

Current controls include:

• A bachelor's in computer science, mathematics, or a similar field
• Strong preparation in mathematics and theoretical computer science
• A published German-equivalent minimum grade of 2.9
• Either German C1 plus English B2, or English C1 for an English route
• For the current English-only route, a medium-of-instruction confirmation is not sufficient
• APS by the application deadline for applicants with Indian qualifications
• Baden-Wuerttemberg tuition for many non-EU/EEA students, subject to exemptions

This is not a standalone "Machine Learning and Computer Vision MSc." Confirm the modules required to earn the AI specialisation and whether the project and thesis can be completed in your intended area.

Official Freiburg Computer Science MSc page

University of Bonn: Computer Science MSc

Bonn's consecutive Computer Science MSc is a broad research-oriented programme with four research areas, including Graphics/Vision/Audio and Intelligent Systems.

The programme currently requires:

• A relevant university degree
• Prior modules in mathematical foundations, theoretical computer science, programming, software technology, and information systems
• A scientific thesis
• The programme's accepted English evidence

The university's pages currently contain differing English-level summaries, so applicants should use the binding admission regulation and current application checklist rather than a search snippet or third-party table.

This route can support vision through graphics/vision/audio modules and a thesis, but applicants should create a semester-by-semester plan before relying on that track label.

Official Bonn Computer Science MSc page

Bonn Institute programme page

TUM: Informatics MSc

TUM's Informatics MSc is an English, 120 ECTS research programme that builds on undergraduate computer-science fundamentals. Students can form an individual specialisation, but computer vision is not a separate degree admission route.

Current application controls include:

• A two-stage aptitude assessment
• Curriculum analysis
• A scientific essay
• VPD for many degrees obtained outside Germany
• GRE General or GATE evidence for applicants whose qualifying bachelor's was completed in specified countries including India
• APS where applicable
• EUR 6,000-per-semester tuition for many newly enrolled third-country students, subject to TUM's exemptions and waivers

Applicants should inspect the current module catalogue for computer vision, geometry, machine learning, robotics, graphics, and systems, then verify scheduling.

Official TUM Informatics MSc page

TUM: Robotics, Cognition, Intelligence MSc

TUM's Robotics, Cognition, Intelligence MSc is a distinct interdisciplinary route combining informatics, electrical engineering, and mechanical engineering. Its curriculum includes perception, image processing, sensor analysis and fusion, robotics, machine learning, and intelligent autonomous systems.

Important distinctions:

• Both German and English are required.
• Modules are taught in both languages.
• The programme has a two-stage aptitude assessment.
• Current international third-country tuition is EUR 6,000 per semester, subject to exemptions and waivers.
• Computer Vision II: Multiple View Geometry appears in the current cognition curriculum.

Choose this route for robotics perception and autonomous systems only if you also want the controls, sensing, and robotics components. It is not an English-only general vision programme.

Official TUM Robotics, Cognition, Intelligence page

TU Berlin: Computer Science MSc

TU Berlin offers a broad, English-language 120 ECTS Computer Science MSc with winter and summer entry. "Open admission" does not mean no academic requirements.

The current page requires:

• 36 credits in computer-science foundations, including 12 each in theoretical computer science, computer engineering or information technology, and methodological-practical computer science
• 18 credits in mathematics
• At least 30 additional computer-science credits, excluding the thesis
• English B2

TU Berlin warns that fewer than 5% of applicants in recent application periods fulfilled these entry requirements, particularly because theoretical-computer-science preparation was often missing.

Vision must be assembled through the current module offering, projects, and thesis. Do not call this a formal Computer Vision track unless the applicable curriculum says so.

Official TU Berlin Computer Science MSc page

University of Tuebingen: Machine Learning MSc

Tuebingen's English Machine Learning MSc is not a dedicated vision degree, but its flexible curriculum can support computer-vision research and projects.

Current formal requirements include:

• A degree related to machine learning, such as computer science, mathematics, physics, or a related field
• A German-equivalent grade of 2.0 or better
• At least 27 ECTS in mathematics
• At least 18 ECTS in computer science, including at least 6 ECTS in algorithms and data structures
• Programme-specific English evidence

For winter 2026/27, the published application deadline was 30 April 2026. Future applicants must check the next cycle rather than reuse that date.

Official Tuebingen Machine Learning admission page

Programme Comparison Snapshot

The table does not replace current regulations or module handbooks.

Build a Formal Prerequisite Map

Create one row for each requirement:

Use official syllabi. A course named "Artificial Intelligence" does not prove coverage of computer vision, deep learning, or the specific formal category used by the university.

Adjacent Degrees

ECE, electrical engineering, mathematics, physics, mechatronics, and other degrees may be eligible for particular programmes. Eligibility depends on the published related-field rule and formal module content.

Do not assume:

• Signal processing equals computer-science theory
• MATLAB equals programming credits
• A vision project equals an algorithms course
• A high CGPA replaces missing prerequisites
• MOOCs create transcript ECTS

For field changes, use the changing-fields guide.

Audit the Curriculum

Download the current module handbook and answer:

1. Is computer vision compulsory, elective, or absent?
2. Are classical vision and geometry taught, or only deep learning?
3. Can you study machine learning foundations before advanced vision?
4. Are graphics, image processing, robotics, or medical imaging available?
5. Which modules run every semester, annually, or irregularly?
6. Are projects capped or allocated competitively?
7. Can the thesis be supervised in the target area?
8. Are modules and examinations available in your usable language?
9. Can the full degree be completed without timetable conflicts?
10. Does the programme teach deployment and systems, or only models?

Build a complete 120 ECTS plan. A lab website or professor name is not a curriculum.

Test Research Access Separately

Research institutes and university groups can strengthen the environment, but proximity does not guarantee access.

For each group:

• Check whether the professor is currently at the institution.
• Review recent projects and publications, not historic reputation.
• Check current teaching and thesis listings.
• Identify whether master's students can join seminars or projects.
• Check prerequisite modules and application procedures.
• Verify whether supervision is available in your intended semester.
• Keep alternative thesis groups.

A research group can change direction, funding, staff, or capacity during a two-year degree.

Build Technical Evidence That Can Be Inspected

Admissions and hiring are separate, but both benefit from rigorous work.

2D Vision Project

Include:

• Dataset provenance and licence
• Train/validation/test split design
• Baseline
• Metrics appropriate to class balance and task
• Error analysis
• Ablations
• Reproducible environment
• Model and data limitations

3D Vision Project

Include:

• Camera model and calibration
• Coordinate frames
• Geometry assumptions
• Pose or reconstruction evaluation
• Failure cases
• Runtime and memory

Robotics Perception Project

Include:

• Sensor timing and calibration
• Fusion method
• Real-time constraints
• State estimation
• Closed-loop or recorded-data validation
• Safety behaviour when perception fails

Medical Imaging Project

Include:

• De-identification and lawful data use
• Patient-level data splitting
• Leakage controls
• Clinically meaningful evaluation
• Uncertainty and calibration
• External or domain-shift testing
• Clear non-clinical limitations

Deployment Project

Include:

• Inference latency and throughput
• Hardware and precision
• Memory and energy use
• Monitoring and drift
• Versioned data and models
• Rollback or fallback behaviour

Do not present benchmark accuracy without experimental controls.

Thesis and Company Access Are Competitive

A master's thesis may be:

• Internal at a university group
• External at a research institute
• Conducted with a company under university examination rules

An external thesis can require an eligible university examiner, topic approval, intellectual-property terms, confidentiality review, and a suitable opening. Payment and employment status vary.

Do not assume that a company partnership or nearby office creates a thesis pipeline or job conversion.

Before accepting:

• Confirm academic ownership and examination rules.
• Review publication and confidentiality restrictions.
• Clarify data access after the contract ends.
• Confirm compute, hardware, and supervision.
• Define deliverables and evaluation.
• Keep enough time for the written academic work.

Career Research by Role, Not Buzzword

"Computer vision engineer" is only one possible title. Search for:

• Machine learning engineer
• Perception engineer
• Robotics software engineer
• Image-processing engineer
• Computer-vision researcher
• Applied scientist
• Medical image-analysis researcher
• Optical metrology or inspection engineer
• Remote-sensing scientist
• ML platform or inference engineer

For 30 to 50 current vacancies, record:

Use the Federal Employment Agency's Entgeltatlas for the closest established occupation and region. Do not infer a fixed starting salary from the word "AI."

Language Planning

The programme's admission language and the labour market are different controls.

• Saarland Visual Computing requires C1 English and rejects MOI proof.
• TUM RCI requires German and English.
• Freiburg's English route requires C1 English under its current rules.
• TU Berlin currently requires English B2.

Employment language depends on team, domain, safety requirements, customers, production environment, and documentation. German can materially widen options, but no single level guarantees employment.

Cost Controls

Check:

• Programme tuition
• Semester contribution
• Baden-Wuerttemberg or TUM fee rules and exemptions
• Housing near the actual campus
• Health insurance
• Compute or equipment needs
• Conference, field-trip, or project costs
• Expected programme duration

TUM's current EUR 6,000-per-semester third-country tuition and Baden-Wuerttemberg's EUR 1,500 state fee can materially change the comparison.

Use the public-university cost method instead of an old university price table.

Application Control Sheet

Save dated copies of material rules.

Decision Scorecard

Score only programmes that pass hard eligibility and affordability gates.

Do not add points for unsupported rankings or employer names.

Red Flags

Re-check advice that:

• Calls a university's broad CS degree a dedicated CV master's
• Ranks programmes by publication counts without a transparent method
• Uses unofficial Indian CGPA thresholds
• Gives one IELTS range for all programmes
• Says CS, ECE, mathematics, or physics graduates are automatically eligible
• Treats self-study as a substitute for formal credits
• Lists fixed deadlines without applicant category and intake
• Uses outdated TUM tuition
• Calls TUM RCI English-only
• Treats TU Berlin's open admission as unrestricted eligibility
• Guarantees access to a lab, institute, company thesis, or employer
• Publishes generic CV salary ranges
• Describes PhDs as universally funded at one salary
• Claims a thesis normally converts into a job

Frequently Asked Questions

Is there a dedicated computer vision master's in Germany?

Saarland University's Visual Computing MSc is a dedicated visual-computing route. Many other universities offer vision through broader Computer Science, Machine Learning, Robotics, or Data and Computer Science degrees.

Is TUM the best option?

There is no universal best option. TUM Informatics offers flexibility and TUM RCI offers a robotics-perception route, but eligibility, language, tuition, module access, and technical goals determine fit.

Can an ECE graduate apply?

Possibly. Programmes such as Saarland Visual Computing explicitly name electrical engineering among related fields. Broad Computer Science programmes may require formal theory, systems, programming, and other credits that an ECE transcript does not contain.

Do I need deep-learning projects?

Only if the programme requires or evaluates them. A rigorous project can support an application or portfolio, but it does not replace formal admission requirements.

Should I learn classical computer vision?

For many profiles, yes. Geometry, image formation, estimation, calibration, and evaluation remain important in 3D vision, robotics, medical imaging, and industrial systems.

Is Heidelberg's programme called Computer Science?

The current programme is Data and Computer Science MSc. Its module handbook includes vision-related study, but applicants should use the current degree title and regulations.

Is Freiburg's degree a Computer Vision MSc?

No. It is a Computer Science MSc with an Artificial Intelligence specialisation that can include computer vision and related modules.

Is TU Berlin easy to enter because it is open admission?

No. The programme has extensive subject-credit requirements, and TU Berlin states that fewer than 5% of applicants in recent periods fulfilled them.

Does a company thesis guarantee a job?

No. Thesis availability, performance, headcount, language, work authorisation, and hiring conditions are separate.

Is German required?

It depends on the programme and role. TUM RCI requires both German and English, while several other programmes in this guide can be studied in English. Employment requirements vary.

Final Checklist

• Define the intended vision profile.
• Verify formal prior-degree and credit eligibility.
• Separate degree title from elective specialisation.
• Build a full module plan.
• Check module frequency and language.
• Identify several thesis groups.
• Audit research access rather than assuming it.
• Build reproducible technical evidence.
• Test current job requirements by role.
• Verify tuition, exemptions, and total cost.
• Save current regulations and application evidence.
• Maintain programme alternatives across degree architectures.

Related guides:

• AI master's programmes in Germany
• AI versus data science
• Embedded systems programmes
• Changing fields for a German master's
• German university application guide

The strongest computer-vision shortlist is not the list with the most famous labs. It is the list of programmes you can enter, afford, complete, and use to build the precise vision capabilities your target work requires.