AI Rotoscoping: Automatic Video Object Segmentation

Manual rotoscoping of a 4-minute scene is 5,760 frames at 24fps. An experienced specialist covers 30–60 frames per day, turning an episode into 3–6 months of work. We automate this process with AI, reducing turnaround to about a week. Our pipeline uses cutting-edge video segmentation models and temporal smoothing for industrial-grade quality. Over 50 projects processed, including feature films and commercials. Clients save up to 70% on rotoscoping budgets, which can be from $5,000 to $50,000 on large projects.

Application Areas of AI Rotoscoping

AI rotoscoping automatically extracts objects (people, cars, animals) from video using computer vision models. Unlike manual rotoscoping, where each frame is traced by hand, AI models segment the object and track its motion. This is indispensable for post-production: background replacement, adding effects, isolating objects for compositing. The technology works for any scene: from simple (static camera, clear outline) to complex (fast motion, flowing hair, crowds).

How AI Rotoscoping Works

Technical Details

The modern approach is not just frame-by-frame segmentation but temporal mask tracking with consistency. Key tools:

SAM 2 (Segment Anything Model 2) by Meta SAM 2 — purpose-built for video segmentation. You provide a point or bounding box on the first frame, and the model propagates the mask through the entire clip accounting for motion. In practice: accuracy holds for 80–120 frames without additional prompts, beyond that correction is needed. SAM 2's memory module retains context from previous frames.

import torch
from sam2.build_sam import build_sam2_video_predictor

predictor = build_sam2_video_predictor(
    "sam2_hiera_large.yaml",
    "sam2_hiera_large.pt",
    device="cuda"
)

with torch.inference_mode(), torch.autocast("cuda", dtype=torch.bfloat16):
    state = predictor.init_state(video_path="scene_001.mp4")
    # Set a point on the actor in frame 0
    _, _, masks = predictor.add_new_points_or_box(
        state, frame_idx=0, obj_id=1,
        points=[[540, 380]], labels=[1]
    )
    # Propagate mask through the whole video
    for frame_idx, obj_ids, masks in predictor.propagate_in_video(state):
        save_mask(frame_idx, masks)

RVM (Robust Video Matting) — specialized in separating people from backgrounds, operates in real-time (30fps on RTX 3080). Better than SAM 2 for scenes with flowing hair and semi-transparent elements.

Why Mask Flickering Happens and How to Fix It

Note: when segmenting frames independently, the mask flickers — edges jump 2–5 pixels between frames. In final compositing, this looks like a trembling outline. Solution — temporal smoothing:

1. Optical flow consistency: apply RAFT or FlowFormer to compute optical flow between frames; warp the mask from frame N to frame N+1 and average with the model prediction.
2. Post-processing with morphological operations: slight erode/dilate on the mask removes noise; Gaussian blur on edges makes transitions smooth.
3. Alpha matting: instead of a binary mask (0/1), use soft alpha (0..1) on edges — via GuidedFilter or Deep Image Matting.

In practice: SAM 2 without temporal smoothing shows edge flickering on fast motion. After applying RAFT + alpha refinement with ViTMatte, mask edge displacement between frames is under 1.2 px (subpixel stability).

How Temporal Smoothing Eliminates Mask Flickering

Temporal smoothing is not just a filter but a set of techniques ensuring temporal mask consistency. Optical flow (RAFT) transfers the mask from frame to frame, while alpha matting (ViTMatte) adds subpixel precision at edges. The result: a stable mask even on complex scenes with fast motion.

Production Workflow

1. Primary automatic segmentation with SAM 2 / RVM — entire clip.
2. QA: automatic flickering detector (mask variance in a 5-frame sliding window > threshold).
3. Manual correction only on problematic frames — via Silhouette, Mocha Pro, or After Effects Roto Brush.
4. Alpha refinement with ViTMatte for hair and semi-transparent fabrics.
5. Export EXR sequence with alpha channel.

Manual-to-automatic ratio: simple scene — 90/10, complex — 60/40. Budget savings on a feature film can reach hundreds of thousands of rubles. Our AI rotoscoping pipeline uses SAM 2 and RVM for automatic video object segmentation.

Contact us for an accurate estimate of your project. We will select the optimal pipeline and calculate the cost individually. Our team has 10+ years experience in VFX, certified in compositing.

What's Included

• Primary automatic segmentation using the chosen model (SAM 2, RVM).
• QA and flickering detection with a report.
• Manual correction of problematic frames (up to 10% of total volume) — free of charge.
• Alpha refinement for hair, smoke, semi-transparent objects.
• Export in required format (EXR, PNG, MOV with alpha channel).
• Training your team on using the resulting masks (on request).

Timelines are approximate; accurate estimate after reviewing source material. Cost is calculated individually based on scene complexity and required quality. Order a test run of a 1-minute clip to verify quality. Get a consultation for your project — we'll help choose the best approach.

Links: RVM

How Distribution Shift Kills CV Model Metrics in Industry

On a production line, a camera is installed to control product quality. The model is trained on 10,000 labeled images—test accuracy mAP 0.84. Deployed to production, and in the first week it misses 30% of defects. Lighting on the line changes between shifts; distribution shift nullifies the metrics. This is a classic story with computer vision in industry, where pattern recognition fails without proper drift handling.

Our engineers, with experience from 60+ computer vision projects, know how to eliminate such scenarios. We guarantee stable model performance under real conditions.

Object Detection: YOLO, RT-DETR, and Everything in Between

YOLO is the standard for real-time detection. YOLOv8 and YOLOv11 from Ultralytics are the most used versions in production: simple API, active community, built-in validation, and export to ONNX/TensorRT. For tasks with high accuracy requirements and less critical latency, RT-DETR, a transformer-based architecture without NMS, gives better mAP on COCO at comparable speed to YOLOv8l.

A typical mistake when training a detector: dataset of 8000 images, 3 classes, fine-tune YOLOv8m—F1 0.73 on validation. Look at confusion matrix—one class is almost never detected. Cause: imbalance 1:23. Solution: oversampling rare class, focal loss for objectness, augmentations (Mosaic, MixUp disabled for rare class as they "blur" it). Transfer learning is mandatory: pretrained on COCO weights reduces data requirement by 10 times. Fine-tuning on 500–2000 domain images yields a working model in 1–2 days on a single GPU.

For edge deployment: export to ONNX → TensorRT engine. YOLOv8n in TensorRT FP16 on Jetson AGX Orin gives 150+ FPS at P99 latency < 8 ms—3 times faster than ONNX Runtime without TensorRT. On server A10G: 700+ FPS for YOLOv8n in TensorRT INT8.

How Does Fine-Tuning YOLO Help in Pattern Recognition?

Suppose you need to find micro-defects on a metal surface—a task with high resolution and class imbalance. We use YOLOv8m pretrained on COCO and fine-tune on 2000 proprietary images. Apply augmentations Mosaic, MixUp, random perspective. After 200 epochs, mAP 0.5 reaches 0.93. Key techniques:

• Focal loss for the objectness head—reduces contribution of easily classified examples.
• Class-balanced sampling—equalizes representation of rare classes.
• Test Time Augmentation (TTA)—increases recall by 5–7% through averaging over flips and scales.

Get a consultation on architecture selection for your task—contact us.

Segmentation: SAM, Mask R-CNN, and Instance Segmentation

SAM (Segment Anything Model) from Meta changed the approach to segmentation. SAM 2 works with video, supports object tracking across frames—for interactive object selection by point or bbox, it's the best out-of-the-box choice. For production instance segmentation without interactive prompting, Mask R-CNN or YOLOv8-seg are used. YOLOv8-seg trains like a regular detector with additional masks, convenient in the same pipelines. Semantic segmentation (each pixel is a class) uses SegFormer, DeepLabV3+. SegFormer-B5 provides a good balance of accuracy and speed for satellite imagery or medical segmentation.

Case study: cell segmentation on microscopic images. Dataset of 400 images with manual annotation. Training Mask R-CNN on ResNet-50 backbone gave IoU 0.61—poor. Problem: objects (cells) overlap; standard NMS kills overlapping predictions. Solution: switch to cellpose (specialized architecture for biomedical tasks) + soft-NMS. IoU increased to 0.79.

OCR: When Tesseract Fails

Tesseract is a starting point for simple tasks: printed text, good lighting, straight layout. As soon as there are handwritten elements, non-standard fonts, perspective distortions, or multi-column layouts, Tesseract degrades quickly.

PaddleOCR is a production-grade solution: text block detection + recognition + structural analysis. Works out of the box for 80+ languages, including Russian. Supports tables and complex document structures. TrOCR (Microsoft) is a transformer OCR with strong results on handwritten text. For Russian handwritten text, fine-tuning is needed: the base model is trained mostly on Latin script.

What to Do When Tesseract Cannot Handle Pattern Recognition on Documents?

For tasks like "extract data from invoices/contracts/passports," we use LayoutLMv3 or Donut—these models understand document layout, not just text. Integration via Hugging Face Transformers, fine-tuning on 200–500 annotated documents. Typical pipeline:

1. Preprocessing: deskew, denoising, binarization via OpenCV.
2. Text block detection: PaddleOCR detection or CRAFT.
3. Recognition: PaddleOCR recognition or TrOCR.
4. Post-processing: normalization, validation via regex or LLM for structured fields.

For documents with fixed structure, template matching + OCR by coordinates is often more reliable than an end-to-end solution.

Face Recognition: Identification and Verification

Face recognition = detection + alignment + embedding + matching. Each stage matters.

Detection: RetinaFace or InsightFace for accurate face localization and keypoints. MTCNN is older but reliable. Embedding: ArcFace (InsightFace) is state-of-the-art for face recognition embeddings. Models iresnet50/iresnet100 pretrained on MS1MV3 (5M identities). Embedding vector 512 float32, comparison by cosine similarity. Threshold tuning: decision threshold is a critical parameter. At threshold 0.6, typical FPR on LFW benchmark is 0.001, TPR is 0.985. In production, threshold must be calibrated to the real distribution: people in masks, with changed appearance, different lighting conditions. Liveness detection is mandatory: MiniFASNet—lightweight model on CPU; FaceX-Zoo contains several pretrained liveness detectors.

Video Analytics

Video is a sequence of frames plus a temporal dimension. A naive approach—detecting on every frame—is expensive.

Tracking: ByteTrack and BoT-SORT are the standard for multi-object tracking. They work on top of any detector, adding persistent IDs to objects across frames—enabling object counting, motion tracking, velocity.

Optimization: not every frame needs processing. For static scenes, detect every 5–10 frames, with tracking in between. For event detection (person entering a zone), background subtraction (OpenCV MOG2) serves as a lightweight pre-filter before neural detection. Action recognition: SlowFast, VideoMAE for action classification. Heavy models—for production use ONNX export + TensorRT or offline processing.

How to Measure Pattern Recognition Model Quality in Production?

Quality monitoring is key to MLOps. We track:

• Prediction confidence distribution.
• Share of low-confidence predictions (indicator of OOD data).
• Drift of input images via feature distribution (embeddings from backbone).

A drop in average confidence from 0.87 to 0.71 over a week is an early signal of distribution shift. NVIDIA Triton Inference Server recommends tracking these metrics via Prometheus. Our certified engineers set up monitoring and guarantee SLA for inference quality.

Deployment of CV Models

For online inference, we use Triton Inference Server (NVIDIA)—production standard for serving CV models. Supports TensorRT, ONNX, PyTorch, dynamic batching, multiple instances. REST and gRPC API. We guarantee stable operation under load.

Edge deployment: ONNX Runtime on ARM/x86 CPU. TensorFlow Lite for mobile devices. OpenVINO for Intel CPU/GPU/VPU—gives 2–3× speedup on Intel hardware compared to ONNX Runtime. After deployment, we hand over the model with documentation and train personnel.

What Is Included in the Work

Deadlines and Cost

A prototype detector on existing data takes 1–2 weeks. Production system with optimization for target hardware takes 4–8 weeks. Full cycle including data labeling (1000–5000 images) takes 2–4 months. Cost is calculated individually for each task. Typical savings from implementing a quality control system can be significant per production line.

We have been in the market for over 5 years and completed 60+ computer vision projects. We will evaluate your project end-to-end—request a consultation to get a quote and technical proposal.