Blog

3D reconstruction of the heart

Using minimal 2D images from a routine angiography, RSIP Vision’s algorithm tracks the veins and constructs a route map of the arteries, from two different angles. Our sophisticated algorithms applied to computer vision in cardiology reconstruct 3D models of the heart artery tree from 2D images, allowing for improved diagnostic capabilities, lower risk to both patients and physicians, more effective therapy and reduced costs.

Defect Detection in Ceramics

When a tile is manufactured in mass production lines, manual inspection becomes a limiting factor to speed of production. This calls for the development of an automated inspection and defects detection in ceramics material, which RSIP Vision has built for one of its clients, generating dramatic improvements in terms of output quality, waste reductions and loss of labor time, all of which benefit the manufacturer’s image and profits.

Grading and sorting

RSIP Vision develops advanced deep learning software for fast and accurate grading and sorting of agricultural produce. One of the key benefits of this solution is its ability to effectively detect existing features and defects, to predict which items will last longer (and therefore can be shipped far away) and which items should be retained for the local market. Sorting and grading machines based on deep learning yield a consistent performance. They are the state-of-the-art solution we recommend today for applications of this kind.

Retinal images enhancement

Image enhancement of retinal structures has the potential to facilitate diagnosis of several eye diseases. Retinal disease diagnosis and monitoring often requires very delicate analysis

OCR Check Scanner

Due to the different styles of handwritten characters, digitalizing check information can be very challenging. But not enough for RSIP Vision’s engineers, who built software for a client to enable him to do just that: detect different handwriting as well as printed check information and transmit all this data from smartphones to the bank. The result is that bank customers using the app which we developed for our client can deposit checks by simply placing their smartphone over the check.

Pattern Recognition for Lens Markings Detection

Lens marking refers to the placement of temporary and/or permanent marking, semi-visible laser engraving for the use of lens identification and trademarks, and accurate placement

Visual Inspection of Semiconductors

Semiconductors mass production needs exceptional levels of precision. Among the many processes needed to insure quality and reliability of chips, we were asked to verify that silicon chip masks imprint the right information on silicon wafers. This is a complex, multi-process computer vision task, to solve which we produced an advanced algorithmic software and vision system, the key advantage of which resides in the capture of critical errors early on in the semiconductor production process.

OCR Software for Low Contrast Images

It is not uncommon that images needing automatic Optical Character Recognition were originally taken in low light or under direct sunlight. RSIP Vision has developed software which accurately binarizes vehicle number plates even in low contrast images, resulting in quick and accurate transmission of license plate details to the client’s database for further action.

Tear film formation and meibomian glands

Meibomian gland dysfunction detection

Meibomian gland dysfunction is often seen as an early stage of dry eye syndrome. Indeed, Meibomian glands play a significant role in tears production by

Glaucoma Detection

Glaucoma, a high intraocular pressure (IOP) pathology, leading to damage of the optic nerve, can be better detected using deep learning techniques. When it detects the optical disc (the visible section of the optic nerve), the deep learning algorithm helps assess glaucoma in an automated way, starting from the region of interest and providing a reliable probability for the disease, which the physician will use to support both diagnosis and treatment decisions.

Coronary CT Angiography with Deep Learning

The production of 3D CT images of the heart requires a fast image processing technology, applied simultaneously on multiple scanned layers. To automatically separate the different components of the image, our software locates in the images the muscular layer (myocardium) of the heart needed for the rest of the segmentation in coronary CT angiography. Minimum graph cuts is the technique which provides the clearest tracking and the strongest segmentation results.

Automated diagnostic reading of HIV

Human Immunodeficiency Virus is a double-strand RNA retrovirus, which translates into a DNA by reverse transcription in the cell cytoplasm and later integrated into the

Automated lung segmentation - airways and blood vessels

Lung Segmentation Software

RSIP Vision has built a lower respiratory tract segmentation software using advanced image processing algorithms. This region includes both lungs along with their pulmonary vasculature. This lung segmentation software takes advantage of the structure of vascular and capillary tree in the area as an exploratory tool for lungs segmentation, enabling both diagnosis and planning of invasive interventions.