Intrinsic Innovation LLC

Massachusetts Institute of Technology

Boston Polarimetrics, Inc.

Mitsubishi Electric Research Laboratories, Inc.

The General Hospital Corporation

Achuta Kadambi

Kartik Venkataraman

Agastya Kalra

Ramesh Raskar

Vage Taamazyan

Ayush Bhandari

Supreeth Krishna Rao

Christopher Barsi

Petros T Boufounos

Boxin Shi

Rishav Agarwal

Guy Michael Stoppi

Barmak Heshmat Dehkordi

Rajiv Gupta

Adrian Andrew Dorrington

Abhijit Ghosh

Lee Vincent Streeter

Bradley Craig Anderson Brown

Tomohiro Maeda

Hang Zhao

Refael Whyte

Adam Pan

Anton Boykov

James Schiel

A computer-implemented method for computing a prediction on images of a scene includes: receiving one or more polarization raw frames of a scene, the polarization raw frames being captured with a polarizing filter at a different linear polarization angle; extracting one or more first tensors in one or more polarization representation spaces from the polarization raw frames; and computing a prediction regarding one or more optically challenging objects in the scene based on the one or more first tensors in the one or more polarization representation spaces.

Systems and methods for transparent object segmentation using polarization cues

A system for collecting data for training a computer vision model for shape estimation includes: an imaging system configured to capture one or more images; and a processing system including a processor and memory storing instructions that, when executed by the processor, cause the processor to: receive one or more input images from the imaging system; estimate a pose of an object depicted in the one or more images; render a shape estimate from a 3-D model of the object posed in accordance with the pose of the object; and generate a data point of a training dataset, the data point including one or more images based on the one or more input images and a label corresponding to the one or more images, the label including the shape estimate.

Systems and methods for generating and using visual datasets for training computer vision models

A data capture stage includes a frame at least partially surrounding a target object, a rotation device within the frame and configured to selectively rotate the target object, a plurality of cameras coupled to the frame and configured to capture images of the target object from different angles, a sensor coupled to the frame and configured to sense mapping data corresponding to the target object, and an augmentation data generator configured to control a rotation of the rotation device, to control operations of the plurality of cameras and the sensor, and to generate training data based on the images and the mapping data.

Systems and methods for training pose estimators in computer vision

A sensor system includes: a radar system configured to emit a radar beam and receive reflected radar signals from in a field of view of the radar system; a camera system including one or more cameras, at least one camera including a linear polarization filter in an optical axis of the camera, a field of view of the camera system overlapping the field of view of the radar system; and a processing system including a processor and memory, the memory storing instructions that, when executed by the processor, cause the processor to: receive radar data based on the reflected radar signals captured by the radar system; receive polarization raw frames captured by the camera system; and compute a track of a target in the field of view of the camera system and the field of view of the radar system based on the radar data and the polarization raw frames.

Sensor fusion between radar and optically polarized camera

Systems and method for an object from a plurality of objects are disclosed. An image of a scene containing the plurality of objects is obtained, and a segmentation map is generated for the objects in the scene. The shapes of the objects are determined based on the segmentation map. An end effector is adjusted in response to determining the shapes of the objects. The adjusting the end effector includes shaping the end effector according to at least one of the shapes of the objects. The plurality of objects is approached in response to the shaping of the end effector, and one of the plurality of objects is picked with the end effector.

Systems and methods for a vision guided end effector

A method for controlling a robotic system includes: capturing, by an imaging system, one or more images of a scene; computing, by a processing circuit including a processor and memory, one or more instance segmentation masks based on the one or more images, the one or more instance segmentation masks detecting one or more objects in the scene; computing, by the processing circuit, one or more pickability scores for the one or more objects; selecting, by the processing circuit, an object among the one or more objects based on the one or more pickability scores; computing, by the processing circuit, an object picking plan for the selected object; and outputting, by the processing circuit, the object picking plan to a controller configured to control an end effector of a robotic arm to pick the selected object.

Systems and methods for picking objects using 3-D geometry and segmentation

A polarized event camera system includes: an event camera having a field of view centered around an optical axis, the event camera including an image sensor including a plurality of pixels, each pixel of the event camera operating independently and asynchronously and being configured to generate change events based on changes in intensity of light received by the pixel; and a rotatable linear polarizer aligned with the optical axis of the event camera, the rotatable linear polarizer having a polarization axis, the polarization axis of the rotatable linear polarizer being rotatable about the optical axis of the event camera.

Polarization event cameras

A method of performing surface profilometry includes receiving one or more polarization raw frames of a printed layer of a physical object undergoing additive manufacturing, the one or more polarization raw frames being captured at different polarizations by one or more polarization cameras, extracting one or more polarization feature maps in one or more polarization representation spaces from the one or more polarization raw frames, obtaining a coarse layer depth map of the printed layer, generating one or more surface-normal images based on the coarse layer depth map and the one or more polarization feature maps, and generating a 3D reconstruction of the printed layer based on the one or more surface-normal images.

Systems and methods for surface normals sensing with polarization

A stereo camera array system includes: a first camera array at a first viewpoint including: a first camera configured to capture images in a first modality, the first modality being viewpoint-independent; and a second camera configured to capture images in a second modality different from the first modality; and a second camera array at a second viewpoint spaced apart along a first baseline from the first camera array at the first viewpoint, the second camera array including: a first camera configured to capture images in the first modality; and a second camera configured to capture images in the second modality.

Multi-modal and multi-spectral stereo camera arrays

A method for estimating a pose of an object includes: receiving, by a processor, an observed image depicting the object from a viewpoint; computing, by the processor, an instance segmentation map identifying a class of the object depicted in the observed image; loading, by the processor, a 3-D model corresponding to the class of the object; computing, by the processor, a rendered image of the 3-D model in accordance with an initial pose estimate of the object and the viewpoint of the observed image; computing, by the processor, a plurality of dense image-to-object correspondences between the observed image of the object and the 3-D model based on the observed image and the rendered image; and computing, by the processor, the pose of the object based on the dense image-to-object correspondences.

Systems and methods for using computer vision to pick up small objects

Systems and method for expanding a dynamic range associated with an image are disclosed. The method includes capturing an image of a scene via an imaging device using a single exposure. The image of the scene includes a plurality of polarization images corresponding to different angles of polarization, and each of the polarization images comprise a plurality of color channels. The method further includes determining a criterion for each of the plurality of color channels; selecting one color channel of the plurality of color channels based on determining of the criterion; generating a reconstructed image irradiance for the one color channel based on pixels in two or more of the plurality of polarization images obtained for the one color channel; and outputting a reconstructed image with the reconstructed image irradiance.

Systems and methods for high dynamic range image reconstruction

A method for estimating a pose of an object includes: receiving a plurality of images of the object captured from multiple viewpoints with respect to the object; initializing a current pose of the object based on computing an initial estimated pose of the object from at least one of the plurality of images; predicting a plurality of 2-D keypoints associated with the object from each of the plurality of images; and computing an updated pose that minimizes a cost function based on a plurality of differences between the 2-D keypoints and a plurality of 3-D keypoints associated with a 3-D model of the object as arranged in accordance with the current pose, and as projected to each of the viewpoints.

Systems and methods for pose detection and measurement

A method of tracking a pose of an object includes determining an initial pose of the object at a first position, receiving position data and velocity data corresponding to movement of the object to a second position by a moving device, determining an expected pose of the object at the second position based on the position and velocity data and the initial pose, receiving second image data corresponding to the object at the second position from a camera, and determining a refined pose of the object at the second position based on the second image data and the expected pose.

In-hand pose refinement for pick and place automation

A multi-modal sensor system includes: an underlying sensor system; a polarization camera system configured to capture polarization raw frames corresponding to a plurality of different polarization states; and a processing system including a processor and memory, the processing system being configured to control the underlying sensor system and the polarization camera system, the memory storing instructions that, when executed by the processor, cause the processor to: control the underlying sensor system to perform sensing on a scene and the polarization camera system to capture a plurality of polarization raw frames of the scene; extract first tensors in polarization representation spaces based on the plurality of polarization raw frames; and compute a characterization output based on an output of the underlying sensor system and the first tensors in polarization representation spaces.

Systems and methods for augmentation of sensor systems and imaging systems with polarization

An optical system includes: a beam splitter system configured to split an input beam into a plurality of output beams including a first output beam, a second output beam, and a third output beam; a first polarizing filter having a first polarization angle and configured to filter the first output beam to produce a first filtered output beam; a second polarizing filter having a second angle of polarization and configured to filter the second output beam to produce a second filtered output beam; and a third polarizing filter having a third angle of polarization and configured to filter the third output beam to produce a third filtered output beam, the first, second, and third angles of polarization being different from one another.

Growing community of inventors

Cambridge, MA, United States of America

Achuta Kadambi