Best computer vision course in India (2026): an honest comparison

Computer vision courses for Indian learners run from free (NPTEL's IIT courses, with an optional Rs 1,000 certificate exam) to ₹14,900 for Skillfloor's live course, USD 162 on offer at OpenCV University, or USD 999 for Udacity's nanodegree. Skill-Lync's computer vision courses are ₹40,000 each. Almost all of this is online: if you searched for classes near you, we found no classroom course with a public fee. Decide between classical vision and deep learning first.

Checked 18 Nov 2026 · Fees and facts from each provider's own page (linked) · Written by Skill-Lync, one of the options compared

The short answer

  • You want to try computer vision before paying anything: Start with OpenCV's free OpenCV Bootcamp (about 3 hours), then audit IBM's Introduction to Computer Vision and Image Processing on Coursera. Both use Python and OpenCV.
  • You want an IIT name on the certificate for very little: NPTEL's 12-week courses are free to learn, with a Rs 1,000 proctored exam for the certificate: IIT Kharagpur's Computer Vision for classical theory, IIT Hyderabad's Deep Learning for Computer Vision for modern methods. E&ICT Academy, IIT Kanpur's 30-hour recorded course shows Rs. 423.73 plus GST, as a faculty price.
  • You already code in Python and want deep learning for vision: DeepLearning.AI's Convolutional Neural Networks course on Coursera (4 weeks at 10 hours a week) or Udacity's Computer Vision Nanodegree (37 hours, USD 249 a month or USD 999 one-time, with project feedback).
  • You want live classes rather than recordings: Skillfloor's Certified Computer Vision Specialist: ₹14,900 for 6 weeks of live online training, then a 6-week internship. Its "100% placement assistance" is a claim, not a placement rate.
  • You want classical computer vision in Python step by step, with reviewed projects: Skill-Lync's Basics of Computer Vision using Python runs 12 weeks from colour spaces and filters to edges, segmentation, optical flow, feature matching, camera calibration and stereo, and ends with CNNs and YOLO. Its two projects use public road videos. Basics of Computer Vision using Python →

Side by side

Fees as published on 18 Nov 2026; they change often, so confirm on the provider's page before you pay. "Not stated" means the page asks you to enquire.

Provider and programFeeDurationFormatToolsCertificatePlacement supportBest for
Skill-LyncBasics of Computer Vision using Python2 projects, about 21 hours of video₹40,000 incl. all charges12 weeks100% Online: recorded lessons, live doubt sessions, graded projectsPython, OpenCV, TensorFlow, Keras TensorFlow 2Skill-Lync certificate of completionNot included with single coursesPython beginners who want classical computer vision built up week by week: colour spaces, filtering, edge detection, Otsu thresholding, morphology, optical flow, Harris, SIFT and FAST features, camera calibration and epipolar geometry, then LeNet, AlexNet and YOLO; projects on pedestrian tracking and road-video segmentation
Skill-LyncApplying CV for Autonomous Vehicles using Python2 projects, about 21 hours of video₹40,000 incl. all charges12 weeks100% Online: recorded lessons, live doubt sessions, graded projectsMATLAB, Python, TensorFlow, OpenCV, Keras TensorFlow 2Skill-Lync certificate of completionNot included with single coursesPython learners heading for vehicle perception: Canny and Hough line detection, stereo and camera calibration, MobileNet image classification, optical flow, Deep SORT and KLT tracking, segmentation, TensorFlow object detection and an introduction to 3D object detection
Skill-LyncIntroduction to Camera Systems2 projects, about 28 hours of video₹40,000 incl. all charges8 weeks100% Online: recorded lessons, live doubt sessions, graded projectsSee syllabusSkill-Lync certificate of completionNot included with single coursesLearners who want the camera side of vision: CCD and CMOS sensors, RCCB filters, pinhole and fisheye models, lens distortion, intrinsic and extrinsic calibration, stereo depth maps, structure from motion and a basic SLAM project, plus camera interfaces such as GMSL and I2C
NPTEL (IIT Kharagpur)Computer Vision (Prof. Jayanta Mukhopadhyay)Free to enrol and learn; the optional proctored certificate exam costs Rs 100012 weeks (the last run's exam was on 24 October 2026; the next run date is not on this page)Recorded video lectures with weekly assignments; certificate only through an in-person proctored exam at a designated centreNo software named on the page; theory-led: image processing, projective and camera geometry, stereo, features and matching, colour, range images, classification, a closing week on deep networksNPTEL certificate from IIT Kharagpur if you pass the paid examNoneEngineering students who want the maths of classical computer vision (homography, camera and stereo geometry) from an IIT professor, at almost no cost
NPTEL (IIT Hyderabad)Deep Learning for Computer Vision (Prof. Vineeth N Balasubramanian)Free to enrol and learn; the optional proctored certificate exam costs Rs 100012 weeks (the last run's exam was on 18 October 2026; the next run date is not on this page)Recorded video lectures with theory and programming assignments; certificate through an in-person proctored examPython programming assignments; covers CNNs, object detection and segmentation, RNNs, attention and vision transformers, GANs, VAEs, diffusion models and vision-language modelsNPTEL certificate from IIT Hyderabad if you pass the paid examNoneLearners who already know basic machine learning, Python, linear algebra and probability and want modern deep learning for vision; the page recommends a prior ML and deep learning course
E&ICT Academy, IIT Kanpur (EICTA consortium)Deep Learning with Generative AI for Computer VisionRs. 423.73 plus GST, the only price shown, tagged 'Faculty' (it is run as a faculty development programme; its FAQ says students and professionals can also join)30 hours, self-pacedRecorded, self-paced; listed as a faculty development program, though its FAQ says students and professionals can joinTensorFlow and PyTorch (per its FAQ); CNNs, transformers, GANs, autoencoders, image and video quality assessmentCertificate from E&ICT Academy, IIT Kanpur on completionNoneA very cheap, short deep-learning-for-vision overview with an IIT Kanpur academy certificate; not a hands-on OpenCV course
Coursera (IBM)Introduction to Computer Vision and Image ProcessingFree to enrol; the certificate needs a paid Coursera subscription. Included with Coursera Plus, shown at ₹7,499 for a year on an offer banner2 weeks at 10 hours a week (6 modules)Recorded, self-paced, with labs and a final transfer-learning project (traffic sign classification)Python, OpenCV, Pillow; Haar cascade, R-CNN and MobileNet object detectionShareable IBM certificate via Coursera (paid)NonePython learners who want a short, low-cost first course in image processing and object detection
Coursera (DeepLearning.AI)Convolutional Neural Networks (course 4 of the Deep Learning Specialization, Andrew Ng)Free to enrol; the certificate needs a paid Coursera subscription. Included with Coursera Plus, shown at ₹7,499 for a year on an offer banner4 weeks at 10 hours a week (4 modules)Recorded, self-paced, graded programming assignmentsPython, TensorFlow; CNNs, residual networks, transfer learning, object detection, face recognition, neural style transferShareable DeepLearning.AI certificate via Coursera (paid)NoneLearners who know Python and basic neural networks and want the standard deep learning route into vision
OpenCV University (opencv.org)Mastering OpenCV with PythonUSD 162 on offer (list price USD 249); the page shows a waitlist with a 20% discount for waitlisted buyers. The 6-course CVDL Master Program is USD 1,624 on offer (list USD 2,499)21 modules, self-paced with lifetime access (no hours given)Recorded video lessons with code, graded quizzes, assignments and projects; community support; 30-day money-back guaranteePython, OpenCVOpenCV.org certificate of completion (over 50% on graded work)None; the page shows learner testimonials about jobs, not a placement serviceBeginners who want practical OpenCV from the library's own team; no AI or computer vision background needed
UdacityComputer Vision NanodegreeUSD 249 a month, USD 212 a month on a 4-month bundle, or USD 999 one-time (from the page's structured data)37 hoursRecorded, self-paced, projects with expert feedbackPython; CNNs, RNNs and attention, YOLO, SLAM, Kalman filters and feature matchingUdacity Nanodegree certificateCareer coaching and interview prep with the subscription (its claim, no outcomes given)Intermediate Python programmers who want reviewed projects such as facial keypoint detection and image captioning
SkillfloorCertified Computer Vision Specialist₹14,900 (shown against a struck-out ₹60,000; GST not mentioned)6 weeks of live training, then a 6-week internshipLive online classes plus a client-project internshipPython, OpenCV, TensorFlow, PyTorch, YOLO, Jupyter NotebookSkillfloor certificate with NASSCOM FutureSkills Prime recognition (its claim)"100% placement assistance" (its claim, no outcomes given)Learners who want live classes and a short internship at a low fee; check the trainer and the internship project before paying

How to choose

  1. Decide between classical vision and deep learning. Classical computer vision (filters, features, camera geometry, calibration, stereo) still drives camera, robotics, inspection and measurement work. Deep learning drives detection, segmentation and recognition. NPTEL IIT Kharagpur and Skill-Lync's Basics course lean classical; NPTEL IIT Hyderabad, DeepLearning.AI and Udacity lean deep learning. Most jobs expect some of both.
  2. Check the prerequisites honestly. Every course here uses Python. Deep learning courses also assume NumPy, linear algebra and probability: NPTEL IIT Hyderabad recommends a prior machine learning and deep learning course, and Udacity asks for intermediate, object-oriented Python. If that is not you yet, start with a beginner OpenCV course.
  3. Software costs nothing, so do not pay for it. OpenCV, PyTorch, TensorFlow and Keras are free, and free cloud notebooks such as Google Colab give you a GPU for course-sized models. Unlike CAD or CAE, there is no licence to buy. A course fee should pay for teaching, structure and project review, not tools.
  4. Look for projects on real data, with a review. A detector that works on ten clean images says little. Ask which public datasets you train on, whether you report metrics such as accuracy or mAP, and who reviews your code. Udacity's projects include facial keypoints and image captioning; Skill-Lync's Basics projects use public road videos.
  5. Know what the certificate proves. An NPTEL certificate needs a proctored exam at a centre, so it shows you passed a test. Most other certificates show completion. Interviewers in computer vision usually ask about projects you built, so a GitHub repository with two or three working projects matters more than the logo.
  6. Read outcome claims carefully. "100% placement assistance" means help with resumes and interviews, not a guaranteed job. Ask any paid provider for named learners placed in computer vision roles in the last 12 months, and what share of the batch that was.

Skill-Lync is a good fit if

  • You want classical computer vision explained step by step in Python and OpenCV, from filters and edges to feature matching, calibration and stereo, before deep learning.
  • You are heading for automotive perception and want lane lines, tracking, segmentation and TensorFlow object detection, or the camera hardware side: sensors, lens models, calibration and SLAM.
  • You want recorded lessons at your own pace, with doubt support and two graded projects per course.

Pick something else if

  • You only want to try computer vision: OpenCV Bootcamp is free, NPTEL's courses are free to learn and IBM's Coursera course is free to audit. Our Basics course is ₹40,000 for about 21 hours of video.
  • You want the newest methods: our syllabi stop at CNNs, YOLO and GANs, with 3D object detection in the vehicle course. Vision transformers, diffusion and vision-language models are not taught; NPTEL IIT Hyderabad covers them for free.
  • You want PyTorch: our courses use TensorFlow and Keras.
  • You want live classes or a classroom near you: our courses are recorded, with doubt support.
  • You want a job guarantee: we do not offer one, and single courses carry no placement support.

What our Data Science learners were paid

Median starting salary 3.6 LPA (middle half 3 to 4.26 LPA), from 174 Data Science placements that Skill-Lync learners reported, snapshot 15 Nov 2026.

Read this with care: it is a self-selected sample of learners who told us about a job, mostly freshers, and most records are from 2022 and 2023 (by year: 2022 n=51, 2023 n=78, 2024 n=33, 2025 n=12). It is not a placement rate. No provider in this table publishes a comparable audited figure, so ask every one of them, us included, for named recent placements in your target role.

Most frequent employers in those records: Nielseniq (14), Rntbci (10), Kabira Mobility (8), William O'Neil (6), Ola Electric (6), Tcs (5), Ust Global (3), Elite Elevators (3). Full list: recent placements.

Free ways to start first

FAQs

Which is the best computer vision course in India?

It depends on where you start. For free theory with an IIT certificate exam, NPTEL. For a quick practical OpenCV start, OpenCV University or IBM's Coursera course. For deep learning, DeepLearning.AI's CNN course or NPTEL IIT Hyderabad. For live classes, Skillfloor. For a 12-week recorded course from filters to YOLO with two reviewed projects, Skill-Lync's Basics of Computer Vision using Python.

What is the computer vision course fee in India?

As checked on 18 Nov 2026: NPTEL and Coursera courses are free to learn; E&ICT Academy, IIT Kanpur's course shows Rs. 423.73 plus GST (a faculty price); the NPTEL certificate exam is Rs 1,000; Coursera Plus is shown at ₹7,499 a year on an offer; Skillfloor's live course is ₹14,900; Skill-Lync's courses are ₹40,000 each. In US dollars: OpenCV University USD 162 on offer, Udacity USD 249 a month or USD 999 one-time.

Are there computer vision classes near me, or only online?

Mostly online. Computer vision needs only a laptop, Python, free libraries and a free cloud GPU, so nearly every course is taught online. We found no classroom computer vision course with a fee shown publicly. Skillfloor teaches live online, and NPTEL's exam is held at an in-person centre. If a local institute offers classes, ask for its syllabus, trainer and fee in writing.

Can I learn computer vision for free?

Yes, to a solid level. NPTEL's IIT Kharagpur and IIT Hyderabad courses are free to learn, 12 weeks each. Stanford's CS231n notes are public, fast.ai's deep learning course is free, and OpenCV's own bootcamp is free and takes about 3 hours. You pay for certificates, reviewed projects or structure. Many learners should start free and pay later.

What should I learn before a computer vision course?

Python first, then NumPy and basic linear algebra: vectors, matrices and transformations. For deep learning courses, add probability and a first machine learning course; NPTEL IIT Hyderabad recommends both, and Udacity asks for intermediate, object-oriented Python. Beginner courses such as OpenCV University's Mastering OpenCV with Python and Skill-Lync's Basics course start from images and pixels.

What salary does a computer vision fresher get in India?

We have no computer-vision-specific figure. The closest group in Skill-Lync's placement records is data science: median reported starting salary 3.6 LPA, middle half 3.0 to 4.26 LPA, from 174 placements (snapshot 15 Nov 2026). Figures are self-reported, 129 of them from 2022 and 2023. Frequent employers were NielsenIQ, RNTBCI and Kabira Mobility. Treat it as a rough guide, not a promise.

How we made this comparison

We are Skill-Lync, so treat this page as a well-sourced starting point, not a neutral review. Every fact about another provider comes from that provider's own public page on 18 Nov 2026 (linked in the table); we did not use listicles, ads or reviews. Where a page did not publish a fee or duration we wrote "not stated" instead of guessing. Skill-Lync's rows are generated from the same data as our course pages. Henry Harvin's computer vision course shows no fee; we found no computer-vision-specific course pages at Simplilearn, upGrad or Intellipaat, where computer vision sits inside wider AI programs; IIT Roorkee E&ICT's AI program with Masai does not list computer vision; fast.ai, CS231n and Great Learning's free course appear under free resources. If a provider's details have changed or we got something wrong, tell us through the form below and we will correct it.

Other comparisons: Best ADAS course in India · Best ANSYS course in India · Best AutoCAD course in India · Best automotive cybersecurity course in India · Best BIW design course in India · Best BMS course in India · Best CAE course in India · Best CATIA course in India · Best certification courses for mechanical engineers · Best CFD course in India · Best embedded systems course in India · Best EV course in India · Best HyperMesh course in India · LS-DYNA course in India · Best MATLAB course in India · Best PCB design course in India · Pipe rack design course in India · Best Python course for mechanical engineers · Revit and BIM course in India · Best SOLIDWORKS course in India · Best vehicle dynamics course in India · Best wiring harness design course in India · Skill-Lync alternatives · Skill-Lync vs DIYguru · Skill-Lync vs IFS Academy · Skill-Lync vs Intellipaat

Still deciding?

Ask a counsellor what fits your background.

Tell us your branch and the role you want. If one of the other options above suits you better, we will say so.

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.

Talk to a career counsellor

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.