[{"data":1,"prerenderedAt":189},["ShallowReactive",2],{"svc-detail-computer-vision":3,"sub-work-computer-vision":151,"ill-services\u002Fcomputer-vision":188},{"slug":4,"short":5,"title":6,"tagline":7,"answer":8,"summary":9,"icon":10,"subServices":11,"useCases":65,"benefits":81,"process":87,"tools":103,"industries":116,"faq":122,"relatedWork":141,"relatedServices":143,"seo":148},"computer-vision","Computer Vision","Computer Vision development services","See defects, plates, people and products, automatically and at scale.","Computer vision services build systems that detect, recognise and measure what appears in images and video: defects on a production line, licence plates at a gate, people and events on CCTV, products on shelves, findings in medical scans. Manufacturers, retailers, logistics operators and security teams use them to inspect, count and monitor continuously with accuracy that holds up in real conditions.","Object detection, defect detection, face and licence-plate recognition, segmentation, CCTV and video analytics, retail shelf analytics, OCR and visual search, trained on your imagery and deployed to cloud or edge.","eye",[12,16,20,24,29,33,37,41,45,49,53,57,61],{"slug":13,"name":14,"text":15},"object-detection","Object detection & tracking","Detect, count and track objects in images and video streams for operations, safety and analytics.",{"slug":17,"name":18,"text":19},"face-recognition","Face recognition","Identity verification and access-control systems built with consent, privacy controls and liveness checks.",{"slug":21,"name":22,"text":23},"image-segmentation","Image segmentation","Pixel-level segmentation for measurement, medical imaging, agriculture and industrial inspection.",{"slug":25,"name":26,"text":27,"page":28},"defect-detection","Defect detection & quality inspection","Automated visual inspection on production lines that finds scratches, misalignments and missing parts faster and more consistently than manual checks.",true,{"slug":30,"name":31,"text":32},"medical-imaging-ai","Medical imaging AI","Decision-support models for radiology, pathology and dermatology imagery, developed with clinical validation in mind.",{"slug":34,"name":35,"text":36},"cctv-analytics","CCTV & smart surveillance analytics","Intrusion, loitering, PPE compliance, crowd and incident detection over existing camera networks.",{"slug":38,"name":39,"text":40,"page":28},"licence-plate-recognition","Vehicle & licence-plate recognition","ANPR for car parks, gates, tolls and fleets, tuned for regional plate formats, lighting and angles.",{"slug":42,"name":43,"text":44},"retail-shelf-analytics","Retail shelf analytics","Planogram compliance, out-of-stock detection and share-of-shelf from photos or fixed cameras.",{"slug":46,"name":47,"text":48},"pose-estimation","Pose estimation","Human pose and movement analysis for safety, sports, fitness and ergonomics.",{"slug":50,"name":51,"text":52},"drone-vision","Drone & aerial vision","Inspection, mapping and counting from drone and satellite imagery for infrastructure, agriculture and construction.",{"slug":54,"name":55,"text":56},"ocr","OCR & document vision","Read text from scans, photos, labels and forms, including handwriting and low-quality captures.",{"slug":58,"name":59,"text":60},"video-analytics","Video analytics","Event detection, summarisation and search over recorded and live video.",{"slug":62,"name":63,"text":64},"visual-search","Visual search","Search catalogues by image for e-commerce, parts and asset identification.",[66,69,72,75,78],{"title":67,"text":68},"Production-line inspection","Catch defects in real time, reduce rework and returns, and record evidence for every unit.",{"title":70,"text":71},"Access and parking control","Recognise plates and faces to open gates, bill parking and manage fleets without manual checks.",{"title":73,"text":74},"Safety compliance monitoring","Detect missing PPE, restricted-zone entry and unsafe behaviour on existing cameras.",{"title":76,"text":77},"Shelf and inventory monitoring","Know what is out of stock or misplaced across stores from photos and cameras.",{"title":79,"text":80},"Document and label reading","Read shipping labels, meters, forms and IDs into your systems automatically.",[82,83,84,85,86],"Continuous inspection and monitoring at machine speed","Fewer defects, incidents and losses reaching customers","Consistent decisions with recorded evidence","Runs on cloud or edge hardware as required","Models that keep working after launch",[88,91,94,97,100],{"step":89,"text":90},"Feasibility and data review","Assess camera setups, lighting, sample imagery and target accuracy; run a quick baseline before committing.",{"step":92,"text":93},"Data collection and labelling","Capture and annotate representative images, including rare cases, with tooling your team can continue.",{"step":95,"text":96},"Model training and validation","Train detection, segmentation or classification models and validate on held-out, real-condition data.",{"step":98,"text":99},"Deployment","Serve on cloud GPUs or edge devices such as Jetson with the throughput and latency your line or gate needs.",{"step":101,"text":102},"Monitoring and retraining","Track accuracy drift as conditions change and retrain on new examples on a schedule.",[104,105,106,107,108,109,110,111,112,113,114,115],"PyTorch","YOLO","OpenCV","TensorFlow","Detectron2","Segment Anything","NVIDIA Jetson","ONNX Runtime","TensorRT","Roboflow","AWS","Docker",[117,118,119,120,121],"manufacturing","logistics","retail-ecommerce","healthcare","real-estate-construction",[123,126,129,132,135,138],{"q":124,"a":125},"How many images do you need to train a model?","For detection tasks a few hundred well-labelled images per class is often enough for a first model thanks to pre-trained backbones; rare defect types need targeted collection. We start with what you have, run a baseline and tell you where more data is worth gathering. Synthetic augmentation (lighting, rotation, noise) stretches small datasets, but it never fully replaces real examples of the failure modes you care about.",{"q":127,"a":128},"Can it run on our existing cameras?","Usually yes. Most IP cameras can stream to an edge box or server where the models run. We assess resolution, angle and lighting during feasibility and recommend adjustments only when they materially improve accuracy. For older analogue systems an inexpensive encoder gets the feed onto the network; we have yet to find a camera setup that could not be worked with.",{"q":130,"a":131},"Cloud or edge deployment?","Edge devices such as NVIDIA Jetson suit real-time inspection and sites with poor connectivity or privacy constraints; cloud suits batch processing and central analytics. Many deployments do inference at the edge and send events and thumbnails to the cloud. We often start in the cloud for speed of iteration and move to edge once the model and thresholds are stable and the hardware case is proven.",{"q":133,"a":134},"How do you handle privacy for face recognition and CCTV?","We design for the applicable law and your policy: consent and purpose limitation, on-device processing where possible, blurring of non-subjects, retention limits and access logs. Face recognition is only built where there is a lawful, documented basis. For most operational use cases (counting, safety, queue analytics) identity is not needed at all, and we design so it is never captured.",{"q":136,"a":137},"What accuracy can we expect?","It depends on the task, image quality and how the classes are defined. We report precision and recall on your data from a baseline early in the project and set targets together, rather than quoting generic figures. As a rule, well-lit fixed-camera tasks with clear class definitions reach very high accuracy; ambiguous defects and variable outdoor scenes take more data and iteration.",{"q":139,"a":140},"What happens when the environment changes?","Lighting, new products or camera moves can degrade accuracy. We monitor confidence and sample predictions, alert on drift and retrain on new examples; the retraining pipeline is part of the delivery. Your operators can flag wrong predictions from the interface, and those examples feed the next training round, so accuracy improves with use rather than decaying.",[142],"healthcare-data-analysis",[144,145,146,147],"multimodal-ai","data-analytics-predictive-ai","mlops-ai-infrastructure","ai-fine-tuning",{"title":149,"description":150},"Computer Vision Development Services | AIYatch","Computer vision: object and defect detection, licence-plate and face recognition, CCTV and video analytics, retail shelf analytics, OCR and visual search.",[152,170],{"path":153,"title":154,"description":155,"industry":156,"services":157,"results":158,"cover":168,"coverAlt":169},"\u002Fwork\u002Fmanufacturing-defect-detection","Computer-vision defect detection on a packaging line","Camera-based inspection that catches print, seal and label defects in real time on a high-speed packaging line, with results fed to the plant's quality system.","Manufacturing",[4,146,145],[159,162,165],{"value":160,"label":161},"~96%","defect recall on the held-out test set",{"value":163,"label":164},"-60%","defective units reaching final inspection",{"value":166,"label":167},"\u003C40 ms","inference per frame at line speed","\u002Fcovers\u002Fwork\u002Fmanufacturing-defect-detection.svg","Abstract navy cover with a rising cyan course line",{"path":171,"title":172,"description":173,"industry":174,"services":175,"results":177,"cover":187,"coverAlt":169},"\u002Fwork\u002Fmultimodal-product-catalogue","Multimodal AI catalogue system for a home-and-garden retailer","Vision-language models that turn supplier photos and spec sheets into structured, searchable product listings, with attribute extraction, image tagging and visual search.","Retail & E-commerce",[144,4,176],"generative-ai-llm",[178,181,184],{"value":179,"label":180},"~10×","faster new-product listing",{"value":182,"label":183},"~93%","attribute accuracy on the QA sample",{"value":185,"label":186},"+9%","search-to-product-page click-through","\u002Fcovers\u002Fwork\u002Fmultimodal-product-catalogue.svg","\u003Csvg xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"0 0 640 480\" role=\"img\" aria-hidden=\"true\">\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-width=\"5\" stroke-linecap=\"round\" stroke-linejoin=\"round\">\u003Crect x=\"284\" y=\"44\" width=\"72\" height=\"46\" rx=\"10\" stroke-width=\"3.5\"\u002F>\u003Ccircle cx=\"320\" cy=\"67\" r=\"13\" stroke-width=\"3.5\"\u002F>\u003Ccircle cx=\"320\" cy=\"67\" r=\"5\" fill=\"var(--accent, #0E7490)\" stroke=\"none\"\u002F>\u003Cpath d=\"M170 144V112H202M438 112H470V144M470 306V338H438M202 338H170V306\" stroke-width=\"6.5\"\u002F>\u003Crect x=\"240\" y=\"152\" width=\"160\" height=\"148\" rx=\"18\"\u002F>\u003Cline x1=\"240\" y1=\"196\" x2=\"400\" y2=\"196\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003Crect x=\"336\" y=\"232\" width=\"48\" height=\"48\" rx=\"6\" stroke-width=\"3.5\"\u002F>\u003Ccircle cx=\"360\" cy=\"256\" r=\"7\" fill=\"var(--accent, #0E7490)\" stroke=\"none\"\u002F>\u003Cline x1=\"100\" y1=\"372\" x2=\"540\" y2=\"372\" stroke-width=\"3.5\"\u002F>\u003Ccircle cx=\"140\" cy=\"372\" r=\"9\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003Ccircle cx=\"230\" cy=\"372\" r=\"9\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003Ccircle cx=\"320\" cy=\"372\" r=\"9\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003Ccircle cx=\"410\" cy=\"372\" r=\"9\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003Ccircle cx=\"500\" cy=\"372\" r=\"9\" stroke-width=\"3.5\" stroke-opacity=\"0.45\"\u002F>\u003C\u002Fg>\u003C\u002Fsvg>",1792973685740]