Skip to content
AI360Xpert
Learning Paths

Computer Vision Engineer

A comprehensive guide for Computer Vision Engineer

""1PREREQUISITES2IMAGE PROCESSING FUNDAMENTALS3MACHINE LEARNING BASICS4DEEP LEARNING FOR COMPUTER VISION5ADVANCED COMPUTER VISION TASKS63D COMPUTER VISION7VIDEO ANALYSIS8DOMAIN-SPECIFIC APPLICATIONS9DEPLOYMENT AND EDGE AI10RESPONSIBLE AI AND MLOPS FOR CVComputer Vision EngineerMathematicsLinear AlgebraVectorsMatricesEigenvaluesSVDTensorsCalculusDerivativesGradientsPartial DerivativesChain RuleJacobiansProbability andStatisticsDistributionsBayes TheoremHypothesis TestingMAP EstimationOptimizationGradient DescentConvex OptimizationLagrange MultipliersHessianProgrammingPythonNumPySciPyPandasOpenCVPillowScikit-imageC++PointersMemory ManagementSTLOpenCV C++ APICMakeDeep LearningFrameworksPyTorchTensorFlowJAXKerasGPU ProgrammingCUDACuDNNTritonVersion ControlGitGitHubGitLabShell ScriptingBashCLI UtilitiesDigital Image RepresentationPixels and ResolutionColor SpacesRGBHSVYCbCrLabGrayscaleImage FormatsJPEGPNGTIFFRAWWebPImage Sensors and ISPImage SignalProcessing PipelineImage Enhancement and FilteringPoint OperationsHistogram EqualizationThresholdingContrast StretchingGamma CorrectionSpatial FilteringConvolutionBlurGaussianMedian FilterBilateral FilterFrequency DomainFilteringFourier TransformHigh-passLow-pass FiltersWaveletsGeometricTransformationsAffinePerspectiveHomographyWarpingFeature Detection and ExtractionEdge DetectionSobelCannyPrewittLaplacianCorner DetectionHarris CornerShi-TomasiLocal FeatureDescriptorsSIFTSURFORBFASTBRIEFMorphologicalOperationsDilationErosionOpeningClosingSkeletonizationImage Registration andFeature MatchingRANSACFLANNBlob DetectionLoGDoGMSERSupervised LearningRegressionLinearRidgeLassoPolynomialClassificationLogistic RegressionSupport VectorMachinesK-Nearest NeighborsNaive BayesEnsemble MethodsDecision TreesRandom ForestsGradient BoostingXGBoostLightGBMUnsupervised LearningClusteringK-MeansDBSCANHierarchicalClusteringMean ShiftDimensionalityReductionPrincipal ComponentAnalysist-SNEUMAPAutoencodersEvaluation MetricsClassification MetricsAccuracyPrecisionRecallF1-ScoreROC-AUCPR-AUCConfusion MatrixRegression MetricsMSERMSEMAER-SquaredValidation TechniquesCross-ValidationHold-outBootstrappingK-FoldDeep Learning FundamentalsArtificial NeuralNetworksPerceptronsMulti-LayerPerceptronsTraining MechanismsBackpropagationLoss FunctionsOptimizers - AdamAdamWSGDRegularizationDropoutWeight DecayBatch/Layer/GroupNormalizationData AugmentationAttention MechanismsSelf-AttentionCross-AttentionConvolutional Neural Networks (CNNs)Core ConceptsConvolution LayersPooling LayersStridePaddingReceptive FieldDilated ConvolutionsClassic ArchitecturesLeNetAlexNetVGGInception/GoogLeNetResNetDenseNetEfficientNetLightweightArchitecturesMobileNetShuffleNetSqueezeNetGhostNetAdvanced Deep Learning ConceptsTransfer Learning andParameter-EfficientFine-TuningPEFTLoRAQLoRASelf-SupervisedLearningContrastive LearningSimCLRBYOLMoCoMAEDINO/DINOv2Vision TransformersViTSwin TransformerDETRDeiTLarge MultimodalModelsLMMs - CLIPLLaVAFlamingoGPT-4VGeminiObject DetectionRegion-based MethodsR-CNNFast R-CNNFaster R-CNNCascade R-CNNSingle-Shot DetectorsYOLO seriesSSDRetinaNetAnchor-free DetectorsCenterNetFCOSCornerNetTransformer-basedDetectorsDETRDeformable DETREvaluationIntersection overUnion (IoU)Mean Average Precision(mAP)Image SegmentationSemantic SegmentationFCNU-NetDeepLabV3+PSPNetSegFormerInstance SegmentationMask R-CNNYOLACTPointRendPanoptic SegmentationPanoptic FPNDETR for SegmentationMask2FormerFoundation Models forSegmentationSAMSAM 2 - SegmentAnything ModelEvaluation MetricsDice CoefficientPixel AccuracyMean IoUPose Estimation and Keypoint DetectionHuman Pose EstimationOpenPoseHRNetMediapipeHigherHRNetHand and FacialLandmark Detection3D Pose EstimationImage Generation and EditingGenerative AdversarialNetworksDCGANStyleGANCycleGANPix2PixStarGANVariationalAutoencodersVAEsVQ-VAEDiffusion ModelsStable DiffusionDALL-E 3MidjourneyControlNetDDPMDDIMGenerative EvaluationMetricsFIDInception ScoreCLIP ScoreDocument and Text AnalysisOptical CharacterRecognitionTesseractEasyOCRCRAFTCRNNDocument LayoutAnalysisLayoutLMTrOCRDonutZero-shot and Few-shot LearningMeta-LearningMAMLPrototypical NetworksPrompt Tuning andVisual PromptingCamera Geometry and CalibrationPinhole Camera ModelIntrinsic andExtrinsic ParametersCamera CalibrationCheckerboard patterndistortioncoefficientsZhang's MethodStereo Vision and DepthEpipolar GeometryFundamental andEssential MatricesStereo MatchingDisparity mapsBlock matchingSemi-Global MatchingMonocular DepthEstimationMiDaSDepthAnything3D Point Clouds and Representation3D Data FormatsPoint cloudsVoxel gridsMeshesSDFsProcessing PointCloudsPCL libraryOpen3DDeep Learning on PointCloudsPointNetPointNet++VoxelNetKPConv3D Object Detectionand TrackingPointPillarsSECONDCenterPointLidar ProcessingStructure from Motion and SLAMStructure from MotionSfMCOLMAPVisual SLAMORB-SLAMDSOVINS-MonoVisual-InertialOdometryVIONeural 3D RepresentationsNeural Radiance FieldsNeRFMip-NeRFInstant NGP3D Gaussian SplattingMotion EstimationOptical FlowLucas-KanadeHorn-SchunckDense vs SparseRAFTBackground SubtractionGaussian MixtureModelsMOG2Action RecognitionTwo-Stream NetworksSpatial and Temporalstreams3D CNNsC3DI3DSlowFastX3DSkeleton-based ActionRecognitionGraph ConvolutionalNetworks - GCNsST-GCNObject TrackingSingle Object TrackingSiamese NetworksSORTSiamRPNOSTrackMultiple ObjectTrackingDeepSORTByteTrackFairMOTStrongSORTVideo Generation and EditingText-to-VideoDiffusionSoraRunway Gen-2AnimateDiffVideo Super-Resolutionand TemporalInterpolationAutonomous Vehicles and RoboticsSensor FusionCameraLiDARRadarBird's Eye View (BEV)PerceptionLane and ObstacleDetectionMedical Image AnalysisMedical Formats andProcessingDICOMNIfTI3D SegmentationV-Net3D U-NetDisease Detection andDiagnosticsRemote Sensing and Satellite ImageryMultispectral andHyperspectral ImagingSynthetic ApertureRadar (SAR) AnalysisModel OptimizationQuantizationPost-TrainingQuantizationQuantization-AwareTrainingINT8FP16GPTQAWQPruningWeight pruningStructured vsUnstructuredSparsityKnowledge DistillationHardware-Aware NeuralArchitecture SearchNASDeployment FrameworksOpen Neural NetworkExchangeONNXONNX RuntimeNVIDIA TensorRTGPU optimizationIntel OpenVINOCPU/VPU optimizationApache TVM and MLIRDeep LearningCompilersEdge Computing and HardwareEdge DevicesNVIDIA JetsonNano/OrinRaspberry PiGoogle Coral TPUOAK-DMobile DeploymentTFLiteCoreMLNCNNMNNExecuTorchWeb DeploymentTensorFlow.jsWebAssemblyWebGLWebGPUONNX Runtime WebCloud APIsAWS RekognitionGoogle Cloud VisionAzure Computer VisionEthics, Privacy, and RobustnessBias and FairnessDataset BiasDemographic ParityDebiasing TechniquesPrivacy-PreservingVisionFederated LearningFace AnonymizationDifferential PrivacyAdversarial RobustnessAdversarial AttacksDefensesDeepfake DetectionData ManagementData Annotation ToolsCVATLabel StudioRoboflowScale AIData VersioningDVCPachydermLakeFSActive Learning andHuman-in-the-LoopHITLSynthetic DataGenerationOmniverseUnityBlenderCARLAModel Lifecycle ManagementExperiment TrackingMLflowWeights & BiasesNeptune.aiClearMLModel RegistriesMLflow Model RegistrySagemakerVertex AICI/CD and ProductionContainerizationDockerNVIDIA ContainerToolkitOrchestrationKubernetesKubeflowServing InfrastructureTensorFlow ServingTorchServeTriton InferenceServerBentoMLMonitoring andAlertingData Drift DetectionModel PerformanceMonitoringOut-of-DistributionDetectionEdge Device ManagementFleet managementOTA updates