Practical Computer Vision Mastery: 20+ Python & AI Projects

Meet the instructor and explore artificial intelligence from scratch. Understand how AI works, its fields like machine learning, deep learning, NLP and generative AI, and engage with real-world projects.

Explore visual outcomes of computer vision projects before you start, aligning with practical computer vision mastery across Python and AI projects.

Understand how artificial intelligence mirrors human intelligence—from input data to processing and output actions. Explore machine learning, deep learning, NLP, computer vision, data science, generative AI, and the lifecycle.

Explore the evolution of computer vision from edge detection and haar cascades to cnn s, r-cnn, yolo, mask r-cnn, and semantic, instance, and panoptic segmentation, then vision transformers shaping applications.

Uninstall any existing Python versions, install Python 3.12 with the Windows 64-bit installer, add Python to Path, and verify the installation by checking the version in cmd.

Install vs code on your machine and prepare for python development with a guided windows 10/11 setup. Verify the installation by checking the version via cmd.

Explore computer vision fundamentals with OpenCV, and learn to read, resize, apply color conversion, and save images, applying pre-processing, masking, and basic image processing for real-time projects.

Install and verify OpenCV, import cv2, and read and write images using cv2.imread and cv2.imwrite. Practice loading a color image from a file path and saving it to disk.

Learn practical color space conversion with OpenCV, including grayscale creation, single-channel expansion, and BGR/RGB/HSV transformations, plus resizing and compression for computer vision projects.

Learn how to display images and process videos with OpenCV by using cv2 imshow and waitKey, and by capturing from files, integrated or external cameras, or rtsp streams.

Rotate an image by 45 degrees clockwise using a 2d rotation matrix for data augmentation, then crop a region with OpenCV via warpAffine, imread, and imshow.

Draw lines, rectangles, circles, and text on images using opencv's cv2 functions, enabling line crossing analysis for people-counting systems and in/out counting.

Explore OpenCV thresholding to convert an image to binary. See simple and adaptive (gaussian) and Otsu methods for varying lighting and foreground background separation, with a fixed threshold of 127.

Explore OpenCV morphological operations to reduce noise and enhance image structure, including dilation, erosion, morphology open and close, and gradient for edge detection, using NumPy and Python.

Explore OpenCV contours: detect and draw contour curves that outline boundaries, extract x,y data for object detection, segmentation, and shape analysis using findContours and drawContours.

Build gastric cancer image dataset for object segmentation by annotating normal, stage one, and stage two regions with polygon masks, export vg json, and generate masks with python and cv2.

Explore OpenCV background subtraction by implementing Gaussian mixture model and KNN to separate foreground from a static background in video, using a camera feed and visualizing both methods.

Explore OpenCV bitwise operations for image masking and region extraction, using and, or, xor, and not on image pairs. Create images with NumPy, draw shapes, perform operations, and view results.

Explore histogram equalization to improve image contrast in low light, then apply gamma correction using NumPy and OpenCV to enhance brightness and display the gamma corrected output.

Apply Gaussian blur to reduce noise and blur while preserving edges, then explore median blur for salt-and-pepper noise and the bilateral filter for edge-preserving smoothing.

Apply the laplacian filter to sharpen images and highlight edges by detecting rapid intensity changes, then implement unsharp masking using a gaussian blur to enhance sharpness.

Explore edge detection with the Sobel operator, computing horizontal and vertical intensity gradients, and compare it with Prewitt and Roberts operators using practical Python code samples.

Discover how image processing enhances deep learning and computer vision by improving image quality. Apply contrast adjustment with CLAHE on grayscale images using an 8x8 tile grid and clip limit.

Develop advanced pre-processing techniques for real-time computer vision, including noise removal, image enhancement, feature extraction, and techniques for object detection, segmentation, tracking, optical flow, facial recognition, and optical character recognition.

Master canny edge detection with gradient-based thresholding and non-maximum suppression, using thresholds 50 and 150, then detect lines with the line transform in OpenCV.

Explore corner detection in OpenCV using Harris and fast detectors. Implement Harris corner detection with grayscale images in Python OpenCV, and visualize detected corners.

Explore feature detection and matching in OpenCV, including SIFT, SURF, ORB, and FLANN-based matching, and learn how descriptors enable object recognition and image stitching.

Explore OpenCV texture analysis for feature extraction in face recognition, object detection, and segmentation. Learn two methods—the Gabor filter and local binary pattern—and how they reveal textures and edges.

Explore optical flow and motion analysis to estimate motion between frames, enabling object tracking, action recognition, video stabilization, and motion-based video saving with OpenCV methods like Lucas-Kanade, Horn-Schunck, and Farneback.

Master object tracking in computer vision by assigning unique IDs to detected objects with Kalman filter-based prediction and IOU matching, using detectors like YOLO and faster r-cnn in real-time tracking.

Explore image segmentation that masks regions in images using region-growing and the watershed algorithm with OpenCV, including seed points, thresholds, and pixel similarity.

Detect objects in live video with OpenCV, drawing blue bounding boxes around faces, then explore Mediapipe landmarks for hands and faces, including finger counting and volume or mouse control demos.

Display live streaming in a local web browser using Flask and OpenCV, capturing frames with cv2 and streaming via a video feed API to an HTML template.

Explore optical character recognition with Pytesseract to extract license plate text from images, including preprocessing, using contours to locate the plate, and thresholding.

Explore a real-time smart face system project with five modules: enroll faces, extract embeddings, train models, perform recognition, and build a Tkinter-based graphical user interface.

Install the project dependencies with pip from the requirements file, including OpenCV, CMake, NumPy, scikit-learn, Tkinter, and face recognition models, then run the code in VSCode.

Explore phase enrollment module using Tkinter, OpenCV, and face recognition to enroll students, capture up to 30 faces, create ID dataset folders, and store IDs, names, and encodings in JSON.

Explore face embedding extraction and landmark identification, detailing feature extraction, embedding dimension 128, and a Python Tkinter workflow to encode 210 images into a pickle file for face recognition.

Train face embeddings data with a support vector classifier to build a facial recognition model. Load embeddings, encode labels, configure SVC parameters, fit the model, and save as pickle files.

Demonstrates real-time face recognition for attendance using a trained model, encoding faces, classifying them, and recording daily attendance via a Tkinter interface and JSON logs.

Design and implement an attendance management GUI for a face recognition system, enabling enrollment and attendance views, CSV downloads, and ID deletion for dataset maintenance.

Explore a real-time driver drowsiness detection project using computer vision, featuring facial detection and landmark detection. Implement real-time video processing, alert generation, and a Tkinter GUI.

Explore a driver drowsiness detection project that uses eye closure and yawning detection to prevent accidents, and learn how to run the code and count yawns.

Open the driver drowsiness detection project in VSCode, import cv2 and numpy, and explore mediapipe for face landmark extraction focusing on ears and mouth points.

Implement drowsiness detection by computing eye aspect ratio (ear) and mouth aspect ratio (mar) with MediaPipe landmarks, using Euclidean distance to determine open eyes and mouth across five consecutive frames.

Build a Tkinter GUI for driver drowsiness monitoring, displaying face detection status and yawning counts, with buttons to start and stop a live cv2 video stream.

Detect driver drowsiness in real time by analyzing ear and mouth aspect ratios from face landmarks via a live video stream, updating frames and visuals with a Tkinter window.

Install Mediapipe to enable real-time driver drowsiness detection with a live feed, facial landmark detection, and mouth, eyes, and yawning counting via a Tkinter interface.