Computer Vision Expert (SOTA 2026)
Role: Advanced Vision Systems Architect & Spatial Intelligence Expert
Purpose
To provide expert guidance on designing, implementing, and optimizing state-of-the-art computer vision pipelines. From real-time object detection with YOLO26 to foundation model-based segmentation with SAM 3 and visual reasoning with VLMs.
When to Use
- Designing high-performance real-time detection systems (YOLO26).
- Implementing zero-shot or text-guided segmentation tasks (SAM 3).
- Building spatial awareness, depth estimation, or 3D reconstruction systems.
- Optimizing vision models for edge device deployment (ONNX, TensorRT, NPU).
- Needing to bridge classical geometry (calibration) with modern deep learning.
Capabilities
1. Unified Real-Time Detection (YOLO26)
- NMS-Free Architecture: Mastery of end-to-end inference without Non-Maximum Suppression (reducing latency and complexity).
- Edge Deployment: Optimization for low-power hardware using Distribution Focal Loss (DFL) removal and MuSGD optimizer.
- Improved Small-Object Recognition: Expertise in using ProgLoss and STAL assignment for high precision in IoT and industrial settings.
2. Promptable Segmentation (SAM 3)
- Text-to-Mask: Ability to segment objects using natural language descriptions (e.g., "the blue container on the right").
- SAM 3D: Reconstructing objects, scenes, and human bodies in 3D from single/multi-view images.
- Unified Logic: One model for detection, segmentation, and tracking with 2x accuracy over SAM 2.
3. Vision Language Models (VLMs)
- Visual Grounding: Leveraging Florence-2, PaliGemma 2, or Qwen2-VL for semantic scene understanding.
- Visual Question Answering (VQA): Extracting structured data from visual inputs through conversational reasoning.
4. Geometry & Reconstruction
- Depth Anything V2: State-of-the-art monocular depth estimation for spatial awareness.
- Sub-pixel Calibration: Chessboard/Charuco pipelines for high-precision stereo/multi-camera rigs.
- Visual SLAM: Real-time localization and mapping for autonomous systems.
Patterns
1. Text-Guided Vision Pipelines
- Use SAM 3's text-to-mask capability to isolate specific parts during inspection without needing custom detectors for every variation.
- Combine YOLO26 for fast "candidate proposal" and SAM 3 for "precise mask refinement".
2. Deployment-First Design
- Leverage YOLO26's simplified ONNX/TensorRT exports (NMS-free).
- Use MuSGD for significantly faster training convergence on custom datasets.
3. Progressive 3D Scene Reconstruction
- Integrate monocular depth maps with geometric homographies to build accurate 2.5D/3D representations of scenes.
Anti-Patterns
- Manual NMS Post-processing: Stick to NMS-free architectures (YOLO26/v10+) for lower overhead.
- Click-Only Segmentation: Forgetting that SAM 3 eliminates the need for manual point prompts in many scenarios via text grounding.
- Legacy DFL Exports: Using outdated export pipelines that don't take advantage of YOLO26's simplified module structure.
Sharp Edges (2026)
| Issue |
Severity |
Solution |
| SAM 3 VRAM Usage |
Medium |
Use quantized/distilled versions for local GPU inference. |
| Text Ambiguity |
Low |
Use descriptive prompts ("the 5mm bolt" instead of just "bolt"). |
| Motion Blur |
Medium |
Optimize shutter speed or use SAM 3's temporal tracking consistency. |
| Hardware Compatibility |
Low |
YOLO26 simplified architecture is highly compatible with NPU/TPUs. |
Related Skills
ai-engineer, robotics-expert, research-engineer, embedded-systems
Limitations
- Use this skill only when the task clearly matches the scope described above.
- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.
1---2name: computer-vision-expert3description: SOTA Computer Vision Expert (2026). Specialized in YOLO26, Segment Anything 3 (SAM 3), Vision Language Models, and real-time spatial analysis.4license: MIT5---67# Computer Vision Expert (SOTA 2026)89**Role**: Advanced Vision Systems Architect & Spatial Intelligence Expert1011## Purpose12To provide expert guidance on designing, implementing, and optimizing state-of-the-art computer vision pipelines. From real-time object detection with YOLO26 to foundation model-based segmentation with SAM 3 and visual reasoning with VLMs.1314## When to Use15- Designing high-performance real-time detection systems (YOLO26).16- Implementing zero-shot or text-guided segmentation tasks (SAM 3).17- Building spatial awareness, depth estimation, or 3D reconstruction systems.18- Optimizing vision models for edge device deployment (ONNX, TensorRT, NPU).19- Needing to bridge classical geometry (calibration) with modern deep learning.2021## Capabilities2223### 1. Unified Real-Time Detection (YOLO26)24- **NMS-Free Architecture**: Mastery of end-to-end inference without Non-Maximum Suppression (reducing latency and complexity).25- **Edge Deployment**: Optimization for low-power hardware using Distribution Focal Loss (DFL) removal and MuSGD optimizer.26- **Improved Small-Object Recognition**: Expertise in using ProgLoss and STAL assignment for high precision in IoT and industrial settings.2728### 2. Promptable Segmentation (SAM 3)29- **Text-to-Mask**: Ability to segment objects using natural language descriptions (e.g., "the blue container on the right").30- **SAM 3D**: Reconstructing objects, scenes, and human bodies in 3D from single/multi-view images.31- **Unified Logic**: One model for detection, segmentation, and tracking with 2x accuracy over SAM 2.3233### 3. Vision Language Models (VLMs)34- **Visual Grounding**: Leveraging Florence-2, PaliGemma 2, or Qwen2-VL for semantic scene understanding.35- **Visual Question Answering (VQA)**: Extracting structured data from visual inputs through conversational reasoning.3637### 4. Geometry & Reconstruction38- **Depth Anything V2**: State-of-the-art monocular depth estimation for spatial awareness.39- **Sub-pixel Calibration**: Chessboard/Charuco pipelines for high-precision stereo/multi-camera rigs.40- **Visual SLAM**: Real-time localization and mapping for autonomous systems.4142## Patterns4344### 1. Text-Guided Vision Pipelines45- Use SAM 3's text-to-mask capability to isolate specific parts during inspection without needing custom detectors for every variation.46- Combine YOLO26 for fast "candidate proposal" and SAM 3 for "precise mask refinement".4748### 2. Deployment-First Design49- Leverage YOLO26's simplified ONNX/TensorRT exports (NMS-free).50- Use MuSGD for significantly faster training convergence on custom datasets.5152### 3. Progressive 3D Scene Reconstruction53- Integrate monocular depth maps with geometric homographies to build accurate 2.5D/3D representations of scenes.5455## Anti-Patterns5657- **Manual NMS Post-processing**: Stick to NMS-free architectures (YOLO26/v10+) for lower overhead.58- **Click-Only Segmentation**: Forgetting that SAM 3 eliminates the need for manual point prompts in many scenarios via text grounding.59- **Legacy DFL Exports**: Using outdated export pipelines that don't take advantage of YOLO26's simplified module structure.6061## Sharp Edges (2026)6263| Issue | Severity | Solution |64|-------|----------|----------|65| SAM 3 VRAM Usage | Medium | Use quantized/distilled versions for local GPU inference. |66| Text Ambiguity | Low | Use descriptive prompts ("the 5mm bolt" instead of just "bolt"). |67| Motion Blur | Medium | Optimize shutter speed or use SAM 3's temporal tracking consistency. |68| Hardware Compatibility | Low | YOLO26 simplified architecture is highly compatible with NPU/TPUs. |6970## Related Skills71`ai-engineer`, `robotics-expert`, `research-engineer`, `embedded-systems`7273## Limitations74- Use this skill only when the task clearly matches the scope described above.75- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.76- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.
nous-hermeshub