Biosense, Inc.

Nec Laboratories America, Inc.

Carmel—haifa University Economic Corp Ltd.

Margarita Osadchy

Daniel Reisfeld

Matthew Lynn Miller

Binyamin Pinkas

Yann Andre LeCun

Ayman Jarrous

Alexander A Goldin

Boaz Moskovich

A method for privacy-retaining face identification comprising dividing at least one input face into a first set of patches, providing an assortment of patch profiles for each patch in the first set of patches, each patch profile in the assortment being associated with a unique index thereby to define an assortment of unique indices, for each individual patch in the input face, finding at least one patch profile within the assortment of patch profiles which most resembles the individual patch, thereby to define a first set of indices from among the assortment of unique indices whose corresponding patch profiles most resemble the input face's patches respectively, and using a privacy preserving computation to generate a physical output indicative of a comparison of the input face and at least one target face by comparing first and second functions of the first set of indices and of a second set of indices respectively, the second set of indices corresponding to patch profiles which most resemble a second set of patches into which the target face has been divided.

System for secure face identification (SCIFI) and methods useful in conjunction therewith

A method for human face detection that detects faces independently of their particular poses and simultaneously estimates those poses. Our method exhibits an immunity to variations in skin color, eyeglasses, facial hair, lighting, scale and facial expressions, and others. In operation, we train a convolutional neural network to map face images to points on a face manifold, and non-face images to points far away from that manifold, wherein that manifold is parameterized by facial pose. Conceptually, we view a pose parameter as a latent variable, which may be inferred through an energy-minimization process. To train systems based upon our inventive method, we derive a new type of discriminative loss function that is tailored to such detection tasks. Our method enables a multi-view detector that can detect faces in a variety of poses, for example, looking left or right (yaw axis), up or down (pitch axis), or tilting left or right (roll axis). Systems employing our method are highly-reliable, run at near real time (5 frames per second on conventional hardware), and is robust against variations in yaw (&#xb1;90&#xb0;), roll(&#xb1;45&#xb0;), and pitch(&#xb1;60&#xb0;).

Synergistic face detection and pose estimation with energy-based models

A condition, such as an electrical property, of a chamber of a heart is surveyed with a catheter having a condition sensor at its distal tip. A method for surveying the chamber includes the steps of acquiring a first image of the chamber containing topographical information, advancing the distal tip of the catheter into the chamber, acquiring a second image including a representation of the catheter distal tip in the chamber, displaying a superposition of the topographical information extracted from the first image with the second image to generate a displayed superimposed image, acquiring condition information at a point on the chamber displayed on the superimposed image on or near the topographical information, and acquiring condition information at one or more additional points sufficient in number and spacing to permit the generation of a survey map of the condition in the chamber. The method preferably further includes marking the display of the superimposed image to identify the points on the chamber at which the condition information was acquired.

Method and apparatus for intracardially surveying a condition of a chamber of a heart

A method for intracardially mapping a condition such as an electrical or mechanical property of a chamber of a heart uses a catheter having a distal tip and at least one condition sensor contained therein or proximate thereto. The at least one sensor is capable of sensing condition information in the chamber and provides the three-dimensional position of the catheter tip in a positional frame of reference. The method includes the steps of acquiring first and second images of the chamber. The images are acquired from different projections and contain topological information of the chamber. The chamber images are registered with the positional frame of reference. The catheter distal tip is advanced into the catheter and is navigated to acquisition points under the guidance of topological information contained in or derived from the images. Condition and position information are acquired at each of the acquisition points, the points being sufficient in number and spacing to permit the generation of a map of the condition in the chamber. The topological information used to guide the navigation of the catheter is preferably a three-dimensional reconstruction of the chamber derived from the topological information contained in the images.

Growing community of inventors

Haifa, Israel

Margarita Osadchy