Height×width×channels

Height×width×channels is a common notation used to describe the dimensions of image data, particularly in the fields of computer vision and deep learning. It represents the shape of a single image or a batch of images. The first dimension, height, refers to the number of pixels vertically in the image. The second dimension, width, refers to the number of pixels horizontally. The third dimension, channels, indicates the number of color planes or intensity values for each pixel.

For grayscale images, the channel dimension is typically 1, representing a single intensity value per pixel. Color images commonly use 3 channels, corresponding to the red, green, and blue (RGB) color model. Other color spaces like HSV or YUV might also use 3 channels but with different interpretations. In some advanced applications, images might have more than 3 channels, for instance, with the inclusion of an alpha channel for transparency or infrared spectrum data.

This notation is fundamental for understanding how image data is structured and processed by algorithms. Deep learning models, for example, often expect input data in this specific height×width×channels format. When working with datasets or model architectures, it is crucial to be aware of these dimensions to ensure compatibility and correct data handling. Libraries and frameworks frequently use this convention when defining image tensors.