Grayscale image
A grayscale image represents each pixel with a single achromatic intensity value rather than separate color components.
Learn Generative AI with PyTorch
Images represented only by intensity values rather than color channels, used here as a simpler target for generation.
- How it is used
- Uses single-channel images as approachable inputs and targets for classification, autoencoders, and generative models.
- How deep it goes
- Explains one-channel structure, 784-pixel dimensions, preprocessing, visualization, and dense GAN generation, while deferring high-resolution image modeling to CNNs.
- Examples
- Uses Fashion-MNIST clothing items, MNIST handwritten digits, 28 × 28 layouts, and a 784-output GAN generator.
- What you can do after
- Apply via PyTorch and Torchvision by loading, normalizing, reshaping, visualizing, classifying, or generating grayscale images.
Read how Learn Generative AI with PyTorch explains grayscale image →
How Computers Make Books
Grayscale is a range of achromatic intensities between black and white.
- How it is used
- Uses intermediate gray tones to smooth low-resolution screen shapes and reproduce pictures with limited black-and-white media.
- How deep it goes
- Explains gray-based smoothing and names reproduction methods from thresholding through digital halftoning and Floyd–Steinberg dithering.
- Examples
- Compares low-resolution lines and a filled shape in black and white versus gray shades, alongside historical photograph reproduction.
- What you can do after
- Evaluate/understand how gray shades can make low-resolution edges smoother and simulate tones using limited media.
Read how How Computers Make Books explains grayscale image →
Build Your Own Robot: Using Python, CRICKIT, and Raspberry Pi
Grayscale is an image representation in which each pixel records a single intensity value rather than separate color components.
- How it is used
- Prepares camera images for face detection by reducing color pixels to intensity values.
- How deep it goes
- Defines single-channel intensities, shows image shapes and conversion code, then combines conversion with histogram equalization.
- Examples
- Uses a 640-by-480 face photograph, converts it with cv2.cvtColor, and marks the detected face on the original frame.
- What you can do after
- Apply it via the OpenCV cv2 library to prepare images for a face classifier.
Read how Build Your Own Robot: Using Python, CRICKIT, and Raspberry Pi explains grayscale image →
Related concepts
- cvtColorproduced by · OpenCV cvtColor converts an image between color representations such as color and grayscale.
- thresholdused by · A threshold is a boundary value against which a quantity is compared to determine whether a condition is met or an action or decision is triggered.
- digital halftoningaddressed by · Digital halftoning is the conversion of continuous-tone image information into spatial patterns of discrete dots or binary pixels that approximate intermediate tones.
- Fashion-MNISTused by · Fashion-MNIST is a labeled dataset of grayscale images representing ten categories of clothing and accessories for machine-learning training and evaluation.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.