Fordel StudiosVision systems built for production conditions, not lab conditions.
A vision model that scores well on a held-out split of your training data and fails on your production line is not a production system — it is a benchmark artifact. We build systems validated against the actual distribution they will encounter: lighting variation, occlusion, camera angles, and the edge cases that only show up in production.
Vision model failures in production are almost always distribution failures. The training data was collected under controlled conditions. The production environment has lighting variation, camera angle variance, partial occlusion, and image quality fluctuations that were not represented in training. The model learned the easy signal from the controlled data and cannot generalize to the harder signal in real conditions.
The tools are mature and not the bottleneck. YOLOv8 (Ultralytics) achieves real-time detection at high frame rates on CPU hardware. Detectron2 (Meta Research) is the standard for instance segmentation. OpenCV handles the preprocessing and geometric transformation work that turns raw camera input into model-ready tensors. INT8 quantization with ONNX or TensorRT makes deployment on edge hardware viable without prohibitive accuracy trade-offs. What fails is the data pipeline design, the augmentation strategy, and the evaluation methodology — not the model architecture.
- Training data collected under controlled conditions that do not represent deployment
- No monitoring for distribution shift — degradation is invisible until users report failures
- Inference hardware mismatch — GPU assumed in architecture, CPU available in production
- No confidence threshold — model returns wrong answer instead of deferring to human review
- Post-processing logic that works on benchmark images but breaks on real edge cases
- Augmentation strategy that does not cover the actual variation axes in the deployment environment
We start every vision project with a domain analysis: what inputs will the model see in production, what are the variation axes (lighting, angle, distance, occlusion, image quality), and how representative is the existing training data against those axes. If the data does not cover the production distribution, we design a collection strategy and augmentation pipeline before training starts.
Model selection is driven by deployment constraints, not benchmark leaderboards. YOLOv8 is the right choice when real-time processing is required on commodity hardware. Detectron2 is appropriate when segmentation accuracy matters more than throughput. For quality inspection with fine-grained defects, we evaluate whether a classification model on crops outperforms a detection model on full images — the answer is task-specific.
Vision system build process
Analyze existing data against the production variation axes. Identify distribution gaps. Design augmentation strategy — lighting, angle, occlusion, noise — to synthetically cover gaps before collection or training.
Select architecture based on latency budget, hardware constraints, and accuracy requirements. Document the trade-off explicitly: the fastest model that meets accuracy requirements, not the most accurate model available.
Fine-tune on domain-specific data with augmentation. Evaluate against a test set that includes production-representative conditions. Report per-class metrics — aggregate accuracy hides class imbalance problems.
Package model for target hardware with ONNX export, TensorRT for GPU, OpenVINO for Intel edge hardware, or CoreML for Apple Silicon. Apply INT8 quantization where latency or memory constraints require it, with accuracy verification against the eval set.
Instrument production inference for confidence score distribution and prediction class distribution. Drift from established baselines triggers review before user-visible failures accumulate.
Production-distribution augmentation
We design augmentation pipelines that synthetically cover the actual variation axes in your deployment environment: lighting conditions, camera angles, distance variation, partial occlusion, and image quality degradation. This improves generalization without requiring exhaustive manual data collection.
YOLOv8 and Detectron2 fine-tuning
We fine-tune pre-trained YOLOv8 and Detectron2 checkpoints on domain-specific data. Fine-tuning dramatically outperforms zero-shot application of general models for domain-specific tasks like industrial defect detection or specialized object categories.
Hardware-appropriate optimization
INT8 and FP16 quantization for deployment hardware constraints. ONNX export for cross-platform compatibility. TensorRT for NVIDIA GPU deployment. OpenVINO for Intel edge hardware. CoreML for Apple Silicon. The model runs on your hardware at your latency requirement — not on the hardware we developed it on.
Confidence-based human escalation
Production vision systems need a "not sure" output path. We implement confidence thresholds that route low-confidence detections to human review queues. The model handles the easy cases automatically and defers the hard cases to human judgment.
Production drift monitoring
We instrument confidence score distributions and prediction class distributions in production. Drift from established baselines triggers automated alerts. Degradation is caught before failures accumulate to user-visible levels.
- Domain analysis report with production distribution assessment and collection/augmentation strategy
- Fine-tuned model with per-class metrics against production-representative test set
- Inference pipeline with preprocessing, confidence scoring, and human escalation routing
- Deployment package for target hardware with quantization configuration
- Integration with downstream processing pipeline or storage system
- Production monitoring dashboard with confidence distribution and drift detection
Vision systems built with production-representative training data and proper confidence routing handle high-confidence cases autonomously, reduce human review to genuinely ambiguous cases, and maintain quality through drift monitoring that catches distribution shift before failures compound.
Common questions about this service.
How much training data do we need?
It depends on the task, the model architecture, and how similar your domain is to the pre-training data. Object detection on classes similar to COCO categories can achieve strong results with hundreds of annotated images per class using transfer learning. Highly specialized domains — specific industrial components, proprietary document types — require more. We assess data requirements during the domain analysis phase and give realistic estimates before committing to a timeline.
Can vision models run on edge devices?
Yes, with appropriate model selection and optimization. YOLOv8 Nano runs at real-time frame rates on Raspberry Pi-class hardware. INT8 quantized models run on Coral Edge TPU. ONNX Runtime enables cross-platform deployment. The trade-off is accuracy — smaller, faster models accept lower mAP. We quantify the accuracy/latency trade-off for your specific task and hardware.
How do you handle privacy with camera-based systems?
Privacy-preserving techniques proportional to the sensitivity: on-device inference (images never leave the device), anonymization of sensitive regions before storage, data retention policies, and access controls on annotation platforms. For consumer-facing deployments, we recommend reviewing applicable privacy regulations in your jurisdiction before finalizing system design.
Detection, segmentation, or classification — which do we need?
Classification answers "what is in this image." Detection answers "where are the objects and what class are they" using bounding boxes. Segmentation answers "which pixels belong to each object" using masks. Detection is the most common starting point for production systems. Segmentation is necessary when precise object boundaries matter. Classification is appropriate for image-level decisions — pass/fail quality gates, scene categorization.
Ready to get started?
Tell us what you are building. We will scope it, price it honestly, and give you a clear plan.
Start a ConversationFree 30-minute scoping call. No obligation.