Image recognition


Image recognition

SAGE is a platform for optimizing the learning of computers and people. It includes NLU, Text recognition and ATS (automatic simplification of text). Computers learn to read and understand (humanly) documents, data, content and images.

All images that come from digital cameras (including IR, X-ray, terahertz) or scanners   are bitmap files. A bitmap is literally a map of bits that form a particular picture when rendered to a display (monitor). Bitmap files are made up of pixels in a grid. When a bitmap displays a colored image, there are several shades of gradation in colors and lighting. Each pixel in the bitmap might have 16, 24, or 48 bits of information associated with it. The more bits, the greater the resolution of the bitmap – and the larger the file. JPG and GIF files are types of compressed bitmap files. SAGE automatically removes  noises , repairs and enhances images, expands and reduces the file size to achieve optimal  recognition and shrink the file size of the bitmap while maintaining and improving all data that is necessary to render a good image. SAGE automatically zooms in on a bitmap image and adds, deletes, or changes the color and characteristics of individual pixels according to recognition rules.  SAGE automatically links images with associated information and increases the usability of information across the enterprise. Indexing is done automatically.

SAGE supports automatic image repair and enhancement , including numerous filters to “clean-up” images and remove “noises”- small black marks on an image that can come from transmission errors, dirty cameras (scanners ,copiers, etc). It detects the skew of the image by analyzing the text or image elements and modifies the image to remove the skew on binary, gray and color images.  SAGE ability to use both binary and color processes helps to improve image clarity and fidelity. SAGE crops black edges around images, removes the black over important regions, crops the image edges to a definable margin around the text or image, removes halftone backgrounds from an image, while leaving the text alone, finds and removes horizontal and vertical lines and repairs overlapping text and graphics on images, reconstruct horizontal and vertical lines, “redrawing” them straight with smooth edges, finds and removes the marks left on a binary image that have  holes punched in it , removes bumps and spurs on characters and fills in small holes, improves the appearance of an image,  helps reduce compressed file size without losing image detail, increase characters to provide a higher quality image and provides more accuracy for processes such as image recognition, OCR/ICR/Barcode while minimizing costly and typically inefficient manual “Reject/Repair” processes. SAGE can also reduce file sizes, resulting in image files that are both easier to send and more efficient to store.

SAGE works with all kinds of cameras and scanners. It  dramatically improves image quality. SAGE can display images content in relative shades of gray. Images are much easier to understand.  It has central, full-function images control, automated image clean up, images retrieval,  image processing and recognition and communication delivery methods. 

SAGE offers exceptional display performance and functionality as well as multi page viewing and sorting over the Internet. Automatic optimization of image compression and decompression and image caching and prefetching enable optimized recognition and  ensure that even large images take less time to display and less space to store. Users and robots have the ability to easily move, insert, or delete images. They see the files in an intuitive Tree View with thumbnail images of each page. Behind the scenes, SAGE automatically detects standard file naming schemes and reorders, renames, and displays the files correctly according to the user’s actions. Users can use annotations to emphasize what’s important and de-emphasize what’s not; draw attention to objects. SAGE performs geometric and topological detection to facilitate advanced and automated object recognition and segmentation. SAGE can accurately and reliably identify numerous objects in numerous environments. New applications can be developed in plain English and within a short period of time.

Leave your thought here

Your email address will not be published.