Glossary

AVIF A cutting-edge format from the AV1 codec. Surpasses JPEG and WebP in compression with HDR support. Aspect Ratio The width-to-height ratio of images or video. Essential for maintaining shape during resize and crop. Alpha Channel A per-pixel transparency data channel. Enables semi-transparency and smooth cutouts in PNG and WebP. Anti-aliasing A rendering technique that smooths jagged edges (staircase artifacts) on diagonal lines and curves by blending intermediate colors at boundaries. Affine Transformation A linear geometric transformation combining translation, rotation, scaling, and shearing that preserves parallel lines while deforming image geometry - a fundamental image processing operation. Attention Mechanism A neural network component that dynamically computes relevance scores across input elements, enabling the model to focus on the most informative parts of the data. Ambient Occlusion A rendering technique that simulates how ambient light is blocked in crevices, corners, and contact areas between objects, adding subtle shadows that enhance depth perception and spatial realism. Activation Function A non-linear function applied to each neuron's output in a neural network, enabling the model to learn complex patterns beyond linear transformations. Alt Text Descriptive text assigned to images via the HTML alt attribute, providing content information when images cannot be displayed or for screen reader users.

BMP The native Windows raster format storing pixel data uncompressed. Fast to read/write but produces very large files. Brightness The perceived luminous intensity of each pixel. A fundamental quantity used in display calibration and image tonal adjustments. Blend Mode A mathematical method for compositing pixel values from two layers. Includes modes like Multiply, Screen, and Overlay for varied visual effects. Bézier Curve A smooth mathematical curve defined by control points. The foundational technology behind fonts, vector graphics, and animation easing functions. Batch Normalization A technique that normalizes layer inputs across a mini-batch to zero mean and unit variance, stabilizing and accelerating deep network training. Bump Map A grayscale texture technique that perturbs surface normals based on height differences between adjacent pixels, creating the illusion of surface detail through shading without modifying actual geometry. Bokeh The aesthetic quality of out-of-focus blur in an image. Derived from the Japanese word for blur, it evaluates how pleasing defocused areas appear rather than simply measuring blur amount. Backpropagation An algorithm that efficiently computes gradients of the loss function with respect to each network parameter by propagating errors backward from the output layer to the input layer using the chain rule. Blob Binary Large Object - an immutable object representing raw binary data in the browser, widely used for reading and writing image files in JavaScript. Bit Depth The number of bits used to represent color information per pixel. Higher bit depth enables more colors and smoother gradients. Batch Processing Applying identical operations to multiple images simultaneously, enabling efficient bulk resizing, compression, and format conversion. Background Removal The process of separating a foreground subject from its background and replacing the background with transparency or a solid color. Bulk Processing Applying the same operation (resize, compress, convert, etc.) to multiple image files simultaneously, automating repetitive tasks. Brightness Adjustment The process of uniformly increasing or decreasing the luminance of all pixels in an image to correct overall lightness. Backlight A shooting condition where the light source is behind the subject, often causing the subject to appear dark or silhouetted. Blemish Removal Image processing technique for reducing the visibility of skin imperfections such as acne, spots, and pores in portrait photographs. Background Blur An image processing technique that applies blur to the background while keeping the subject sharp, simulating the bokeh effect of wide-aperture lenses. Border A line drawn around images or UI elements, used for visual separation, decoration, or emphasis. Before/After A presentation technique that displays images side by side or with a slider to visually compare the state before and after editing.

Color Depth The number of bits representing color per pixel. Higher bit depth enables smoother gradients and wider tonal range. Color Space A coordinate system for representing colors numerically. sRGB, Display P3, and Adobe RGB define different gamuts. Contrast The difference in luminance between the brightest and darkest areas of an image. Determines the perceived depth, dimensionality, and visual impact of imagery. Clipping Path A closed vector path that defines the visible area of an image, hiding everything outside. Used in print production and web CSS for precise cutouts. Convolution A fundamental image processing operation that slides a kernel (small matrix) across an image, computing weighted sums at each position for blurring, edge detection, and sharpening. CNN (Convolutional Neural Network) A deep learning architecture that uses convolutional and pooling layers to hierarchically extract spatial features from images, becoming the standard model for image recognition. Color Model A mathematical framework for representing colors as numerical values. Multiple models such as RGB, CMYK, and HSV exist for different applications in digital imaging and print. CMYK A subtractive color model using cyan, magenta, yellow, and black inks to reproduce colors on paper. It is the standard for commercial printing. Color Grading The creative manipulation of color in video and photography to establish mood, emotion, and visual storytelling. It is performed after color correction as an artistic process. Chroma Key A compositing technique that treats a specific color (typically green or blue) as transparent, allowing the background to be replaced with different footage or imagery. Compositing The process of combining multiple image or video elements into a single final image, using layer blending, masking, and color correction to achieve a seamless result. Chromatic Aberration An optical defect where a lens focuses different wavelengths of light at different points, causing color fringing or misregistration. Classified into longitudinal and lateral types. Crop Removing unwanted areas from an image to isolate the desired region, commonly used for composition adjustment and aspect ratio correction. Compression Ratio The ratio between original and compressed data sizes. Higher ratios yield smaller files but may introduce quality loss in lossy formats. Canvas API A JavaScript API for pixel-level drawing and image manipulation on HTML5 canvas elements directly in the browser. Chroma Subsampling A compression technique that reduces chrominance channel resolution relative to luminance, exploiting human vision's lower sensitivity to color detail. Color Profile A data file describing the color reproduction characteristics of a device or color space, enabling consistent color appearance across different devices. CDN A geographically distributed network of servers that caches and delivers content from the nearest edge location, reducing latency for image delivery. Cache Busting A technique for forcing browsers and CDNs to fetch updated files by appending version information to filenames or query parameters. Color Picker A UI component or tool that allows users to select and extract color values from images or color palettes. Client-Side Processing Executing image operations entirely within the user's browser without uploading data to a server, ensuring privacy and reducing latency. Color Correction The process of adjusting image colors to achieve natural tones or fix color casts caused by lighting conditions. Camera Shake Image blur caused by camera movement during exposure, resulting in the entire image appearing smeared in one direction. Collage A creative technique of arranging multiple photos or images on a single canvas to form a new composite work. Clip Art Pre-made illustrations and graphic elements designed for insertion into documents, presentations, and web pages. Creative Commons A standardized licensing system that allows creators to specify usage permissions for their works, widely used for image assets and media.

DPI A resolution unit measuring dots per inch. Primary metric for print quality; irrelevant for web display. Diffusion Model A class of generative models that learn to reverse a gradual noising process, generating data by iteratively denoising from pure noise. Data Augmentation A technique that artificially increases training data diversity by applying transformations such as rotation, flipping, and color jittering, improving model generalization and reducing overfitting. Dropout A regularization technique that randomly deactivates neurons during training, preventing co-adaptation and reducing overfitting by implicitly training an ensemble of sub-networks. Downsampling The process of reducing image resolution by decreasing pixel count while applying appropriate filtering to prevent aliasing and maintain visual quality. Displacement Map A grayscale texture that physically moves mesh vertices along their normals based on brightness values, creating real geometric detail that affects silhouettes, shadows, and occlusion. Dynamic Range The ratio between the darkest and brightest luminance levels an image or device can represent. A wider dynamic range captures high-contrast scenes without clipping highlights or crushing shadows. Depth of Field The range of distance in a scene that appears acceptably sharp. Determined by aperture, focal length, and subject distance - shallow DoF isolates subjects while deep DoF keeps entire scenes in focus. Disparity Map An image recording the horizontal pixel displacement (disparity) between corresponding points in a stereo image pair. Depth can be computed from disparity via triangulation, making it an intermediate representation for 3D reconstruction. Data URL A scheme for embedding binary data as Base64-encoded strings within URLs, commonly used to inline small images in HTML or CSS without additional HTTP requests. Drag & Drop A UI pattern that allows users to load files by dragging them from the desktop directly into a browser drop zone, bypassing file dialogs. Download The process of receiving and saving files from a server or cloud storage to a local device.

EXIF A metadata standard embedded in images by cameras and smartphones. Contains shooting parameters and GPS data. Encoder-Decoder A neural network architecture consisting of an encoder that compresses input into a compact latent representation and a decoder that reconstructs the desired output from that representation. Epoch One complete pass through the entire training dataset during model training. Epoch count serves as a fundamental measure of training progress and convergence. EXIF Data Metadata automatically embedded in image files by cameras and smartphones, recording shooting conditions, device info, and GPS coordinates. Exposure The total amount of light reaching the camera sensor, determined by the combination of aperture, shutter speed, and ISO sensitivity.

Feature Point A locally distinctive point in an image, such as a corner or edge intersection, that can be reliably detected and matched across different views for image alignment and tracking. Fourier Transform A mathematical technique that decomposes an image into spatial frequency components, where low frequencies represent smooth regions and high frequencies correspond to edges and noise. File Format A specification defining how image data is structured and stored as binary, determining compression method, color depth, and metadata handling. File Size The amount of storage space an image file occupies, measured in bytes (B), kilobytes (KB), or megabytes (MB). File Extension The suffix after the dot in a filename that identifies the file type and helps the operating system determine how to handle it. Flip Mirroring an image horizontally or vertically by reversing the pixel arrangement along an axis. Filter A general term for processes that apply specific effects to an image, including blur, sharpen, sepia, and mosaic among many others. Focus The state where the camera lens correctly converges light from the subject, producing a sharp and clear image. Fill An operation that fills a designated area with a solid color, gradient, or pattern, commonly known as the paint bucket tool. Favicon A small icon image displayed in browser tabs, bookmarks, and history, serving as a visual identifier for a website.

GIF A lossless indexed-color format supporting up to 256 colors with frame animation, widely used for short looping clips. Gamma An exponent defining the nonlinear relationship between input signal values and output luminance. Gamma correction ensures natural brightness perception on displays. Gradient A smooth transition between two or more colours. Available in linear, radial, and conic varieties for backgrounds, masks, and artistic effects. GAN (Generative Adversarial Network) A deep learning framework that trains two networks - a generator and a discriminator - in an adversarial manner to produce images indistinguishable from real ones. Gaussian Blur A filter that convolves an image with a Gaussian kernel to produce smooth, natural-looking blur by attenuating high-frequency detail. GPS Tag Location data (latitude, longitude, altitude) stored in EXIF metadata, recording where a photograph was taken. Also called a geotag. Grayscale An image format that contains only brightness information, representing tones from black to white without any color data. Gradient Fill A smooth transition between two or more colors, widely used for backgrounds, overlays, and design elements.

HEIC Apple's default image container using HEVC codec. Achieves roughly half the file size of JPEG at equal quality. Histogram A graphical representation showing the distribution of pixel intensities in an image. Used to objectively assess exposure, contrast, and tonal range. Hue The attribute of color that distinguishes red from blue, green from yellow - expressed as an angle (0-360 degrees) on the color wheel representing the dominant wavelength. Homography A 3×3 matrix representing the projective transformation between two planes, used in panorama stitching, perspective correction, and AR marker recognition to describe geometric correspondences. HSV A color model that represents colors using hue, saturation, and value axes. It aligns with human color perception and is widely used in color pickers and image adjustment tools. Harris Corner Detection An algorithm for detecting corners in images by analyzing the eigenvalues of the local gradient autocorrelation matrix, identifying points with significant intensity changes in two directions. Hex Color Code A color notation using six hexadecimal digits to represent RGB channel values, prefixed with #. The standard color format in web development. HDR Photo A photography technique that combines multiple exposures to capture scenes with extreme brightness differences without losing detail in highlights or shadows.

Interpolation Mathematical methods for estimating new pixel values during image scaling, rotation, or warping. Balances quality against computational cost. Image Classification The fundamental computer vision task of assigning a predefined category label to an entire input image, which drove the deep learning revolution and underpins many downstream vision tasks. ICC Profile A data file describing a device's color reproduction characteristics. ICC profiles are the core technology enabling consistent color across monitors, printers, and cameras. Image Pyramid A multi-resolution data structure built by progressively downsampling an image. Used to achieve scale invariance in object detection and template matching. Inference The process of feeding new data into a trained model to obtain predictions. Unlike training, inference does not update model parameters. Image Optimization A set of techniques for minimizing file size while maintaining acceptable visual quality, directly improving web performance and Core Web Vitals. Image Sprite A technique combining multiple small images into a single file, using CSS background-position to display specific regions, reducing HTTP request count. Image Compression Techniques for reducing image file size, broadly categorized into lossless (reversible) and lossy (irreversible) methods. Image Quality A composite measure of visual fidelity encompassing resolution, color depth, compression level, noise, and sharpness. Image Upscaling The process of increasing an image's pixel dimensions beyond its original resolution using interpolation or AI-based super-resolution. Image Format Conversion The process of re-encoding an image from one file format to another, typically for compatibility, size reduction, or feature requirements. Image Encoder A software module that compresses raw pixel data into a specific image format (JPEG, PNG, WebP, etc.) with configurable quality and compression settings. Image Decoder A software module that reads compressed image files and decompresses them into raw pixel data (RGB/RGBA buffer) for display or further processing. Image Size A measure of image dimensions, referring either to pixel dimensions (width x height) or file size (data volume in KB/MB). Image Editor Software applications used to modify digital images through operations such as cropping, color correction, compositing, and retouching. Image Map An HTML technique that defines multiple clickable regions within a single image, each linking to a different destination. Image Search A search feature that retrieves relevant images based on text keywords, with Google Images being the most widely used service.

JPEG A lossy format optimized for photographs. Uses DCT for high compression but degrades on re-compression.

Kernel A small numerical matrix used in convolution operations. The kernel's values determine the filter type - blur, edge detection, sharpening, or emboss. Keying The general process of generating a mask (key signal) to extract specific elements from video or images. Methods include chroma key, luminance key, and difference key.

Lossy Compression A compression method that discards imperceptible data for high compression ratios. Cannot restore originals. Lossless Compression A compression method restoring original data bit-for-bit. Reduces file size with zero quality loss. Latent Space A compressed, lower-dimensional representation space where generative models encode the essential features of high-dimensional data such as images. Lab Color Space A perceptually uniform color space where numerical distance between two colors correlates with perceived difference. It serves as the device-independent connection space in ICC color management. Lens Distortion Geometric aberration where straight lines in a scene are rendered as curves due to lens optical properties. Barrel and pincushion distortion are the two primary types, most pronounced in wide-angle lenses. Lossless A compression method that preserves original data perfectly upon decompression. No quality degradation occurs, but file size reduction is less than lossy compression. Lossy A compression method that discards perceptually insignificant information to achieve substantial file size reduction. Original data cannot be perfectly reconstructed. Lazy Loading A technique that defers loading of off-screen images until the user scrolls near them, significantly improving initial page load speed. Lossless Format An image file format that preserves all original pixel data through compression, allowing perfect reconstruction upon decoding. Layer A system that separates an image into multiple transparent sheets stacked on top of each other, allowing independent editing of each level.

Mask A mechanism that controls visibility of image regions using greyscale values. Layer masks enable non-destructive editing by hiding rather than deleting pixels. Mipmap A pre-computed set of progressively lower-resolution versions of a texture, used by GPUs to reduce aliasing and improve rendering performance when textures are viewed at reduced sizes. Motion Blur The streaking or smearing of an image caused by subject or camera movement during the exposure period. Appears as directional blur along the axis of motion. Metadata Supplementary data embedded in image files describing capture settings, copyright, location, and other properties. Standards include Exif, IPTC, and XMP. Mosaic Filter An image filter that averages pixel blocks to obscure details in selected regions, commonly used for privacy protection. Megapixel A unit representing one million pixels, used to indicate the total number of pixels a camera sensor can capture.

Noise Random variations in brightness or color unrelated to the actual scene content. Becomes prominent in high-ISO photography and long exposures. Neural Network A mathematical model inspired by biological neural circuits, composed of input, hidden, and output layers, serving as the foundation for advanced image processing tasks. Normal Map A texture that encodes surface normal directions as RGB color values, enabling per-pixel lighting calculations that simulate geometric detail on low-polygon models in real-time rendering.

Opacity A value indicating how much a layer or element obscures the content beneath it. Ranges from 0% (fully transparent) to 100% (fully opaque). Object Detection A computer vision task that identifies and localizes multiple objects within an image by predicting bounding boxes and class labels simultaneously. OCR (Optical Character Recognition) Technology that converts text within images or scanned documents into machine-readable digital text, enabling document digitization and automated data extraction. Optical Flow A technique that represents the apparent motion of each pixel between consecutive frames as a vector field. A foundational technology for video analysis, motion detection, and video stabilization. Overfitting A phenomenon where a model learns training data too well, including noise and idiosyncrasies, resulting in degraded performance on unseen data (poor generalization). Orientation Tag An EXIF metadata field (Tag 0x0112) that specifies how image pixels should be rotated or flipped for correct display orientation. OGP Image A thumbnail image automatically displayed when a URL is shared on social media, specified through the Open Graph Protocol meta tags.

PNG A lossless format with full transparency support. Ideal for sharp-edged images like logos and screenshots. Pixel The smallest addressable element of a digital image. Each pixel holds color data, and their aggregate forms the picture. Point Cloud A data format consisting of numerous 3D coordinate points (x, y, z) in space. Acquired by LiDAR or stereo cameras, it serves as foundational data for 3D modeling, autonomous driving, and surveying. Pooling A downsampling operation that reduces the spatial dimensions of feature maps by aggregating values within local regions, lowering computation while adding translation invariance. Progressive JPEG A JPEG encoding method that renders the entire image in progressively sharper passes, allowing users to see a low-resolution preview before full loading completes. PDF A document format developed by Adobe that faithfully preserves text, images, and vector graphics across all platforms and devices. Passport Photo A standardized portrait photograph used for passports, IDs, and official documents, subject to strict size, background, and facial proportion requirements. PPI Pixels Per Inch - a unit measuring display pixel density, indicating how many pixels fit within one inch of screen space. Preview A temporary display feature that shows editing or conversion results before final output, allowing users to verify changes. Panorama A wide-angle image created by stitching multiple overlapping photos together, typically used for landscape and interior photography. Photo Retouching Post-processing work on photographs to improve their appearance by removing blemishes, adjusting colors, and eliminating unwanted elements.

Quantization An optimization technique that represents model weights and activations in lower bit precision (e.g., INT8) to accelerate inference and reduce memory footprint. QR Code A two-dimensional barcode that encodes information such as URLs or text into a square pattern of black and white modules, readable by smartphone cameras.

RAW Unprocessed sensor data from a digital camera. Enables full post-capture control over exposure and white balance. Resolution A measure of image detail expressed as pixel dimensions (width x height) or pixel density (PPI/DPI) by context. Raster Image A digital image composed of a rectangular grid of pixels, where each pixel stores individual color data. Ideal for photographs and complex color gradients. Ray Tracing A rendering technique that traces the path of light rays from the camera through each pixel, simulating reflection, refraction, and shadows with physical accuracy to produce photorealistic images. RGB An additive color model that combines red, green, and blue light to reproduce colors. It is the foundation of color representation on displays, cameras, and web content. Rotoscoping The technique of manually or semi-automatically tracing mask outlines frame by frame to isolate subjects from their backgrounds. Used as a last resort when automated keying fails. RANSAC Random Sample Consensus. An iterative algorithm for robustly estimating model parameters from data containing outliers. Essential for homography estimation and 3D point cloud fitting in computer vision. Regularization A family of techniques that constrain model complexity to prevent overfitting and improve generalization. Weight decay and dropout are the most common examples. Resize Changing the pixel dimensions of an image through interpolation algorithms that compute new pixel values for the target resolution. Responsive Image A technique for delivering optimally sized and formatted images based on the viewing device's screen size and pixel density, preventing unnecessary data transfer. Rotation The process of turning an image by a specified angle around a center point, commonly used to correct orientation. Red-eye Removal A process that detects and corrects the red-eye effect caused by flash photography, restoring natural pupil color. RAW Format A file format that stores unprocessed sensor data from a digital camera, allowing extensive post-processing flexibility without quality loss. Rounded Corners A decorative treatment that rounds the corners of images or UI elements, creating a softer and more approachable visual appearance. Reverse Image Search A search technique that uses an image as the query to find identical or visually similar images across the web.

SVG An XML-based vector image format that scales to any resolution without quality loss. Ideal for icons, logos, and charts. Sharpness The degree of clarity and edge definition in an image. Higher sharpness means more distinct boundaries between adjacent tonal regions. Saturation The intensity or purity of a color. High saturation produces vivid colors, while low saturation approaches gray. Zero saturation yields a completely achromatic tone. Segmentation The process of partitioning an image into meaningful regions (objects, background, parts) by assigning labels to each pixel, a core technique in image analysis. Specular Map A texture that controls the intensity and color of specular (mirror-like) reflections across a 3D model's surface on a per-pixel basis, enabling varied shininess within a single material. Scale Space A continuous framework for representing images at varying levels of detail. By varying the Gaussian blur parameter sigma, image structures can be analyzed uniformly across resolutions. SIFT Scale-Invariant Feature Transform. An algorithm that extracts local image features invariant to scale changes and rotation, serving as a foundational technique for image matching. Stereo Matching A technique that finds corresponding pixels between a pair of images captured from two cameras to recover depth information. A foundational method for 3D measurement in autonomous driving and robotics. Skip Connection A shortcut path in neural networks that bypasses one or more layers, adding or concatenating the input directly to a later layer's output to mitigate vanishing gradients. srcset An HTML img attribute for specifying multiple image candidates, enabling browsers to automatically select the optimal image based on device conditions. Super-Resolution A deep learning technique that infers and generates high-resolution images from low-resolution inputs by reconstructing detail not present in the source. SNS Image Size The recommended image dimensions and aspect ratios for each social media platform, ensuring optimal display without unwanted cropping or recompression. Selection A feature that designates a specific area of an image so that edits apply only within that region. Shadow / Drop Shadow A visual effect that renders a shadow behind images, text, or UI elements to create depth and a floating appearance. Screenshot A feature that captures the current screen display as an image file, commonly used for documentation, sharing, and troubleshooting. Screen Recording A feature that records on-screen activity as a video file, commonly used for creating tutorials, documenting procedures, and capturing gameplay. Stock Photo A service providing professionally shot photographs and illustrations available for licensed or free download for commercial and editorial use.

TIFF A flexible high-quality raster format standard in print and publishing. Supports lossless compression for archival use. Threshold A boundary value used in image processing to classify pixel intensities into binary (black/white) or multiple classes, forming the basis of binarization and segmentation. Transfer Learning A machine learning technique that leverages knowledge from a model pre-trained on a large dataset to improve performance on a different but related task, especially when labeled data is scarce. Texture Mapping A technique for applying 2D images (textures) onto 3D model surfaces to add visual detail such as color, patterns, and material properties using UV coordinate systems. Tone Curve A graph mapping input luminance values to output values. It is the most powerful tonal adjustment tool for precisely controlling brightness, contrast, and tonal distribution in images. Transparency The ability to make portions of an image transparent, essential for creating logos, icons with clear backgrounds, and image compositing. Thumbnail A small-sized version of an image used for previews and listings to help users quickly identify content. Text Overlay A technique of placing text on top of images, commonly used for creating thumbnails, banners, and social media graphics.

Upsampling The process of increasing the resolution of an image or signal by adding pixels and estimating missing information through interpolation or deep learning techniques. Unsharp Mask A sharpening technique that subtracts a blurred copy from the original image to enhance edge contrast and perceived sharpness. Upload The process of sending files from a local device to a server or cloud storage over the internet. Undo/Redo Undo reverses the last action to restore the previous state; Redo re-applies an undone action.

Vector Image A digital image defined by mathematical objects such as points, paths, curves, and fills. Scales to any size without quality loss. Vignetting The darkening of image periphery relative to the center. Occurs naturally due to lens optical properties, or applied intentionally as an artistic effect to draw attention toward the subject. Voxel The smallest unit (volume element) when 3D space is divided into a regular grid. The 3D equivalent of a pixel, used to discretely represent spatial data in medical imaging, 3D modeling, and games.

WebP A next-gen format by Google with both lossy and lossless modes. Better efficiency than JPEG and PNG. Wavelet A technique that decomposes images into multi-resolution representations preserving both spatial and frequency information, serving as the compression basis for JPEG 2000 and effective for denoising. White Point The chromaticity coordinates defining 'white' in a color space. It corresponds to a light source's color temperature and serves as the reference for white balance and color management. Web Worker A browser mechanism for running JavaScript in background threads separate from the main UI thread, preventing interface freezes during heavy image processing. WebAssembly A binary instruction format enabling near-native execution speed in browsers, allowing image processing libraries written in C/C++/Rust to run client-side. Watermark A semi-transparent logo or text overlaid on an image to indicate copyright ownership and deter unauthorized use. White Balance A camera function and image processing technique that adjusts color tones so white objects appear white regardless of the light source.