Course Objective & Scope
The most common visual AI need in the field is locating objects and extracting their boundaries: product counting, defect localisation, vehicle and person detection. The programme compares anchor-based and anchor-free detectors, the YOLO family and the transformer-based RT-DETR, and instance segmentation with Mask R-CNN.
Labelling with CVAT or Label Studio, COCO/YOLO formats, training with Ultralytics, mAP and IoU-based evaluation and tracking with ByteTrack are covered hands-on. The result is a working detection/segmentation prototype for a sample scenario from the organisation.
Who Should Attend?
- AI engineers building detection and counting solutions
- Data scientists working with image data
- Software teams developing camera-based products
- Prerequisite: Python and basic deep learning knowledge
Key Takeaways
- Choose a suitable detector or segmentation model for a task
- Write a labelling guideline and build a dataset in COCO/YOLO format
- Train a custom detection model with Ultralytics
- Evaluate models with mAP, precision-recall curves and IoU thresholds
- Connect detection output to a video stream with tracking
Curriculum
The Detection Problem and Detector Families
- Two-stage and one-stage detectors
- YOLO, RT-DETR and anchor-free approaches
Labelling and Dataset Preparation
- Bounding boxes and polygons with CVAT and Label Studio
- COCO and YOLO formats, label quality control
Training a Custom Detector
- Training with Ultralytics and hyperparameters
- Small objects and class imbalance
Instance Segmentation
- Mask extraction with Mask R-CNN and YOLO-seg
- Mask quality and post-processing
Evaluation and Tracking
- mAP@0.5:0.95, NMS and threshold tuning
- Object tracking and counting with ByteTrack
Other Courses in This Category
-
Computer Vision Bootcamp
Comprehensive Computer Vision Bootcamp ranging from classical image processing techniques to PyTorch-based CNN, Vision Transformer (ViT), Swin Transformer, and ConvNeXt architectures. Hands-on projects cover object detection with YOLO and DETR, segmentation with SAM and U-Net, video analysis, 3D vision (NeRF, Gaussian Splatting), Diffusion models, and model deployment with ONNX/TensorRT.
- Live Course
- 13 modules
-
Semantic, instance and panoptic segmentation: pixel-level training with U-Net, DeepLabV3+, SegFormer, Mask2Former and SAM 2, Dice/mIoU evaluation and mask-based measurement, for industrial, medical and satellite imagery teams.
- Live Course
- 6 modules
-
ViT, DeiT, Swin and DINOv2 architectures: patch embedding, self-attention, comparison with CNNs and fine-tuning with Hugging Face, for engineers selecting and adapting transformer-based models for vision tasks.
- Live Course
- 5 modules
-
Document preprocessing, text recognition, layout analysis and table/form extraction with PaddleOCR, Surya, LayoutLMv3 and Donut, for technical teams turning invoices, contracts and operational documents into structured data.
- Live Course
- 5 modules
-
RTSP video streams, GStreamer/FFmpeg, real-time detection, multi-object tracking with ByteTrack/BoT-SORT, speed/crowd analytics and NVIDIA DeepStream, for teams extracting live operational insights from camera networks.
- Live Course
- 5 modules
-
Human action, ergonomics and anomaly detection with 3D CNNs (SlowFast, X3D), VideoMAE/TimeSformer and skeleton-based models (MMPose, ST-GCN), for teams building workplace safety, assembly-line and behaviour analysis systems.
- Live Course
- 5 modules
Get in touch
