RSIP Vision Learns

Image Features for Classification

Classification problems in image and signal analysis require, on the algorithmic side, to take into account complex information embedded in the data. Images might contain

Segmentation and tracking of kidney stones

An affected anatomical region can be treated in a selective and non-invasive manner by localized and contact-free methods such as high-intensity focused ultrasound. Treatment by

Flexible medical instruments for endoscopic surgery

Instrument tracking for robotic aided surgery

The success of complex fine-scale surgery depends heavily on surgeons’ agility, state of fatigue and their freedom to manipulate miniature instruments in the target organ. Internal organs

Planning of screw insertion trajectory

Robotic assisted minimally invasive surgery has changed the face of surgery. The use of robotic aids in surgery can be traced back to the mid-1980s,

Minimally Invasive Robotic Surgery

Minimally Invasive Surgery (MIS) has emerged as a paradigm shift in modern surgical practices dating back to the early 1990s. What is Minimally Invasive Surgery?

Anatomical surface reconstruction with sparse set of points

Advance navigation techniques in orthopedic surgery allow to construct 3D models of patients’ anatomical parts by matching a surface to an acquired set of data points. As an alternative to costly imaging procedures, ultrasound or fluoroscopic scan can be performed with much reduced costs and hazards. A set of surface points marked by the technician or physician forms then an initial sparse set for anatomical surface reconstruction. Here we show how RSIP Vision’s engineers overcome the challenge of using a sparse set of points to accurately reconstruct a complex 3d surface.

Lesion Detection in CT scan (Hemorrhagic stroke)

Lesion segmentation by random-forest classifiers

Segmentation of lesions in images, such as those obtained from MRI, ultrasound, CT etc, can be viewed as classifying pixels (or voxels, in the 3-D

Gastrointestinal lesion detection with machine learning

Endoscopic examination is performed to disclose the lesion’s biophysical properties and assess its severity according to its physical appearance. Computer vision and machine learning put a large arsenal of techniques at our disposal. Clever utilization of these techniques enables to do just that, that is, translating expert knowledge into a fine-tuned algorithm, specifically designed for the task of GI lesion detection. The highest level of robustness, accuracy, and reproducibility is required, hence automatic methods can perform as a proper alert system for lesion detection.

Tree Detection and Related Applications in Forestry

Using aerial images taken by drone, plane or satellite, RSIP Vision can create forestry image processing and analysis software to efficiently determine: Trees detection Automatic

Forestry Row Detection and Related Applications

RSIP Vision creates forestry image processing and analysis software by using aerial images taken by drone, plane or satellite. Our algorithms enable to efficiently determine:

Point and surface registration in orthopedics

Point and surface registration enable computer vision and image processing to improve surgical orthopedy practices and affect surgery outcome recovery. Bringing point and surface registration in the field of orthopedics, computer vision and image processing hold the potential to improve surgical practices and affect surgery outcome to favor the benefit of patients and fast recovery. Measurement accuracy (within less than 1 mm) is a strict constraint to computer-vision-based algorithms.

Automatic human action recognition in videos

When a video uploader disregards adding tags and categories, online video hosting platforms encounter what is called the new item problem. This can be solved by utilizing visual analysis of videos and images: first, by filtering videos into recognizable objects and combining human action segmentation and recognition; later, by training Multiclass Support Vector Machines to assign labels to detected actions in the temporal domain of videos.

Automatic semantic tagging of images

Sites containing huge amounts of content must necessarily recommend only a narrow and relevant list to users. Recommender systems can be seen as tool to automatically generate personalized search preferences, with the purpose of keeping the balance between monetized targeted suggestions and satisfactory user experience. Automatic semantic tagging helps them do just this.

Automatic video categorization of new item

Online video hosting platforms utilize a variety of methods for content discovery. Recommender systems allow users to face the huge amount of information offered to their view: recommendation algorithms analyze the video and automatically suggest a confined set of adequate tags to the uploader. Alternatively, the system learns to automatically assign tags to videos without any user intervention.

Visual Search and Video Recommendation

Here is a simple explanation of how a Recommender System work. Take YouTube as an example: a huge quantity of videos needs to be processed, classified, tagged and ranked by users or by an automatic algorithm, before it can be used by the Recommender system. Learn how users are classified by filtering (either collaborative, content-based or hybrid) and videos are ranked by similarity.

Vehicle Localization with Lane Tracking

More than 1 million people killed and 50 million injured on the roads every year make driving safety a subject of the highest importance. ADAS technologies are at the forefront of this fight: a further proof of that is our car detection and localization solution enabling to detect a danger, create an alert, avoiding the obstacle or – in the worst case – minimizing damage. This system based on machine learning follows the highest quality standards and can save lives on the road.

Pedestrian Detection with Machine Learning

One of the most challenging tasks of ADAS operating in urban or rural environment is the detection of pedestrians. When human behavior can sometimes be unpredictable, ADAS systems can be programmed to track pedestrians and predict with high levels of accuracy their orientation and intentions: human lives are at stake and our software is key to save them. RSIP Vision‘s algorithms do not save lives only in medical applications, but also in automotive safety systems!

Cardiac Left Ventricle Segmentation

Cardiac MRI is used to assess the status of patients suffering from heart diseases like cardiac masses and thrombi, aortic and/or different kinds of congenital cardiac diseases. Cardiac MRI is minimally invasive, does not involve radiation and it generally delivers excellent images for diagnostics. However, automatic segmentation of the left ventricle on MRI images faces challenging difficulties, like locating the left ventricle and overcoming the lack of edge information. We propose a method providing a very robust input for daily clinical application.

Lesion Border Detection

Dermascopy is one of the main modalities used to detect skin abnormalities and lesions such as malignant melanoma. It is estimated that in the United

Left Ventricle Statistical Shape Modelling

Statistical shape modelling is a powerful tool for visualizing geometric and functional patterns of variation in all organs and also a reliable left ventricle shape model can prove itself very useful, though quite challenging to construct. RSIP Vision knows how to do this and many other image processing solutions for cardiology. And since cardiovascular diseases account for million of deaths per year in the developed world, this work is of vital importance.

Get in touch

Please fill the following form and our experts will be happy to reply to you soon