Computer Vision Bootcamp with Python: YOLO, SAM & RF-DETR

Perform object detection on images using Yolo Nas small with coco weights in a PyCharm project, leveraging Super Gradients and OpenCV to display bounding boxes and confidence scores.

Introduce uranus integration with deepsort for object tracking, test with pre-trained weights, then train on a custom dataset and assign unique IDs to tracked objects.

Learn how to read and display images in Python with OpenCV using cv2.imread, cv2.imshow, and cv2.resize to fit the screen, including installing the OpenCV Python package.

Learn to read, display, and save videos with OpenCV in Python, using VideoWriter to create an output file and test with live webcam capture in the next part.

Capture video from a camera with OpenCV in Python using the internal webcam. Save the live feed to a video file and preview YOLO-based object detection on images and videos.

Track objects in video with the SORT algorithm and YOLO-NAS detection, assigning a unique ID to each object as it enters and leaves the frame.

Learn how to perform object tracking with YOLO-NAS on a custom ship dataset using the Sort algorithm, train on Roboflow data in Colab, and evaluate with mean average precision.

Build a Streamlit app for image object detection using YOLO-NAS, featuring an image uploader, confidence slider, and display of the output image.

Learn to implement object detection on videos and live webcam feeds with a Streamlit app, detailing frame processing, video upload, webcam toggle, and performance considerations.

Merge the object detection code for images and videos into deploy streamlit app.py, then deploy the app on streamlit.io by syncing a GitHub repo with requirements.txt and packages.txt.

Build a streamlit app that counts vehicles entering and leaving by combining object detection with a sort-based tracking algorithm, assigning unique IDs and counting line crossings via center coordinates.

Explore empty shelf detection using YOLO-NAS in a google colab workflow, detailing dataset preparation with roboflow, gpu training, evaluating with mean average precision, and video detections.

Learn how to perform multicam license plate detection and recognition with YOLO-NAS and OCR on multiple real-time video streams, including bounding box visualization and text extraction.

Build a face detection pipeline with yolo-nas in google colab gpu runtime, export data from roboflow, apply augmentation, train, and evaluate with mean average precision at 0.5.

Train and fine-tune a YOLO-NAS model on a PPE dataset to detect seven PPE classes, achieving about 80.8% mAP at 0.5 and testing on images and videos.

Design and implement the home page for a Flask web app that serves video, image, and live webcam detection using YOLO-NAS, wiring templates, routes, and assets.

Train a YOLO-NAS model on a personal protective equipment dataset in Google Colab with a GPU, covering dataset prep from Roboflow, training setup, and detections on images and videos.

Integrate a custom seven-class PPE model into a web app by saving weights, updating class names, and testing with images and videos through a Flask API and a front end.

Train yolo-nas model on a 26-vegetable dataset, balance and augment training data to 19k images, achieving 66.44% mAP on test data, and prototype a streamlit app with chatgpt recipes.

Learn to perform image segmentation with yolo-nas and the segment anything model, producing a person-specific mask from a multi-object output.