# Using Predefined Pipelines

The application includes several predefined pipelines. These pipelines cover common use cases and
can be customized to fit specific requirements.

| Pipeline                                                                                                                     | Description                                                                                                                                                                                                                                                         | Variants                                          |
|------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------|
| ![Age & Gender Recognition](../../_assets/age-and-gender-recognition.png "Age & Gender Recognition")                         | **Age & Gender Recognition**: Retail analytics pipeline using `face-detection-retail-0004` for face detection and `age-gender-recognition-retail-0013` for age and gender prediction. It is suited for customer demographics analysis and similar retail use cases. | Available in CPU, GPU, NPU, and GPU+NPU variants. |
| ![Defect Detection](../../_assets/defect-detection.png "Defect Detection")                                                   | **Defect Detection**: AI-powered pallet defect detection pipeline for manufacturing quality control using machine vision.                                                                                                                                           | Available in CPU, GPU, and NPU variants.          |
| ![Goods Detection](../../_assets/goods-detection.png "Goods Detection")                                                      | **Goods Detection**: Retail pipeline using YOLO 11n object detection to identify retail-related objects for inventory management, customer behavior analysis, and similar use cases.                                                                                | Available in CPU, GPU, and NPU variants.          |
| ![Goods Detection & Classification](../../_assets/goods-detection-and-classification.png "Goods Detection & Classification") | **Goods Detection & Classification**: Retail pipeline using YOLO 11n for detection and EfficientNet B0 for classification to identify and categorize retail-related objects.                                                                                        | Available in CPU, GPU, NPU, and GPU+NPU variants. |
| ![License Plate Recognition](../../_assets/license-plate-recognition.png "License Plate Recognition")                        | **License Plate Recognition**: Simple Video Structurization (D-T-C) pipeline that supports license plate recognition, vehicle detection with attribute classification, and other adaptable detection and classification tasks based on the selected model.          | Available in CPU, GPU, NPU, and GPU+NPU variants. |
| ![Motion Detection](../../_assets/motion-detection.png "Motion Detection")                                                   | **Motion Detection**: Pipeline that uses `gvamotiondetect` to identify motion regions, then runs YOLOv8n object detection restricted to those motion ROIs through `gvadetect`.                                                                                      | Available in CPU, GPU, and NPU variants.          |
| ![Simple NVR](../../_assets/simple-nvr.png "Simple NVR")                                                                     | **Simple NVR**: Lightweight media pipeline for basic video decoding, recording, and format conversion.                                                                                                                                                              | Available in CPU and GPU variants.                |
| ![Smart NVR](../../_assets/smart-nvr.png "Smart NVR")                                                                        | **Smart NVR**: Video analytics pipeline that combines recording with AI-based object detection, tracking, and classification, and produces metadata and processed video frames.                                                                                     | Available in CPU, GPU, NPU, and GPU+NPU variants. |
| ![Smart Parking](../../_assets/smart-parking.png "Smart Parking")                                                            | **Smart Parking**: Cloud-native video analytics pipeline that uses pre-trained deep learning models to detect parking-space occupancy.                                                                                                                              | Available in CPU, GPU, NPU, and GPU+NPU variants. |
| ![Video Summarization VLM](../../_assets/video-summary.png "Video Summarization VLM")                                        | **Video Summarization VLM**: Pipeline using `gvagenai` with a vision-language model to generate concise, scene-level summaries from sampled frames.                                                                                                                 | Available in CPU, GPU, and NPU variants.          |
| ![Segmentation](../../_assets/instance-segmentation.png "Segmentation")                                                      | **Segmentation**: This is a segmentation pipeline preview. Segmentation is used to identify individual objects and separate them from the original image.                                                                                                           | Available in CPU and GPU variants.                |
| ![Human Pose Detection](../../_assets/human-pose.png "Human Pose Detection")                                                 | **Human Pose Detection**: This is a classic physical security use case that detects human poses in real-time. It is particularly useful in use cases such as slip-and-fall detection in elderly care facilities, hospitals, and public spaces.                      | Available in CPU and GPU variants.                |