33 Hits in 0.93 sec

Physics-AI Symbiosis [article]

Bahram Jalali, Achuta Kadambi, Vwani Roychowdhury
2021 arXiv   pre-print
The phenomenal success of physics in explaining nature and designing hardware is predicated on efficient computational models. A universal codebook of physical laws defines the computational rules and a physical system is an interacting ensemble governed by these rules. Led by deep neural networks, artificial intelligence (AI) has introduced an alternate end-to-end data-driven computational framework, with astonishing performance gains in image classification and speech recognition and fueling
more » ... opes for a novel approach to discovering physics itself. These gains, however, come at the expense of interpretability and also computational efficiency; a trend that is on a collision course with the expected end of semiconductor scaling known as the Moore's Law. With focus on photonic applications, this paper argues how an emerging symbiosis of physics and artificial intelligence can overcome such formidable challenges, thereby not only extending the latter's spectacular rise but also transforming the direction of physical science.
arXiv:2109.05959v1 fatcat:gasi2zcd6bhbfd4iq5oqjudlwy

Blending Diverse Physical Priors with Neural Networks [article]

Yunhao Ba, Guangyuan Zhao, Achuta Kadambi
2019 arXiv   pre-print
Machine learning in context of physical systems merits a re-examination of the learning strategy. In addition to data, one can leverage a vast library of physical prior models (e.g. kinematics, fluid flow, etc) to perform more robust inference. The nascent sub-field of physics-based learning (PBL) studies the blending of neural networks with physical priors. While previous PBL algorithms have been applied successfully to specific tasks, it is hard to generalize existing PBL methods to a wide
more » ... ge of physics-based problems. Such generalization would require an architecture that can adapt to variations in the correctness of the physics, or in the quality of training data. No such architecture exists. In this paper, we aim to generalize PBL, by making a first attempt to bring neural architecture search (NAS) to the realm of PBL. We introduce a new method known as physics-based neural architecture search (PhysicsNAS) that is a top-performer across a diverse range of quality in the physical model and the dataset.
arXiv:1910.00201v1 fatcat:ixzodscwu5gutbv3zbaozil7wi

Achieving fairness in medical devices

Achuta Kadambi
2021 Science  
1 2 Fairness Majority performance 1 Pareto inefcient 2 Pareto optimal Reaching the Pareto frontier, where it is not possible to improve fairness without decreasing performance and vice versa Improving fairness is possible to do without changing performance 0-bias Measuring fairness Fairness can be quantified based on e-bias. Fairness is maximized when e = 0, achieving a state of 0-bias. 30
doi:10.1126/science.abe9195 pmid:33795446 fatcat:3ila73phijhzjh2nq6pite4cni

Shape from Mixed Polarization [article]

Vage Taamazyan, Achuta Kadambi, Ramesh Raskar
2016 arXiv   pre-print
Shape from Polarization (SfP) estimates surface normals using photos captured at different polarizer rotations. Fundamentally, the SfP model assumes that light is reflected either diffusely or specularly. However, this model is not valid for many real-world surfaces exhibiting a mixture of diffuse and specular properties. To address this challenge, previous methods have used a sequential solution: first, use an existing algorithm to separate the scene into diffuse and specular components, then
more » ... pply the appropriate SfP model. In this paper, we propose a new method that jointly uses viewpoint and polarization data to holistically separate diffuse and specular components, recover refractive index, and ultimately recover 3D shape. By involving the physics of polarization in the separation process, we demonstrate competitive results with a benchmark method, while recovering additional information (e.g. refractive index).
arXiv:1605.02066v2 fatcat:nmrqaq2wjve4pfukeuzsjz2gfi

Visual Physics: Discovering Physical Laws from Videos [article]

Pradyumna Chari, Chinmay Talegaonkar, Yunhao Ba, Achuta Kadambi
2019 arXiv   pre-print
In this paper, we teach a machine to discover the laws of physics from video streams. We assume no prior knowledge of physics, beyond a temporal stream of bounding boxes. The problem is very difficult because a machine must learn not only a governing equation (e.g. projectile motion) but also the existence of governing parameters (e.g. velocities). We evaluate our ability to discover physical laws on videos of elementary physical phenomena, such as projectile motion or circular motion. These
more » ... mentary tasks have textbook governing equations and enable ground truth verification of our approach.
arXiv:1911.11893v1 fatcat:ilzvb33jjjhgvcisnsgkfbcgrm

Frequency Domain TOF: Encoding Object Depth in Modulation Frequency [article]

Achuta Kadambi, Vage Taamazyan, Suren Jayasuriya, Ramesh Raskar
2015 arXiv   pre-print
Time of flight cameras may emerge as the 3-D sensor of choice. Today, time of flight sensors use phase-based sampling, where the phase delay between emitted and received, high-frequency signals encodes distance. In this paper, we present a new time of flight architecture that relies only on frequency---we refer to this technique as frequency-domain time of flight (FD-TOF). Inspired by optical coherence tomography (OCT), FD-TOF excels when frequency bandwidth is high. With the increasing
more » ... y of TOF sensors, new challenges to time of flight sensing continue to emerge. At high frequencies, FD-TOF offers several potential benefits over phase-based time of flight methods.
arXiv:1503.01804v1 fatcat:i52wi6yurjbpxkxqu27bsblk44

Towards Rotation Invariance in Object Detection [article]

Agastya Kalra, Guy Stoppi, Bradley Brown, Rishav Agarwal, Achuta Kadambi
2021 arXiv   pre-print
Rotation augmentations generally improve a model's invariance/equivariance to rotation - except in object detection. In object detection the shape is not known, therefore rotation creates a label ambiguity. We show that the de-facto method for bounding box label rotation, the Largest Box Method, creates very large labels, leading to poor performance and in many cases worse performance than using no rotation at all. We propose a new method of rotation augmentation that can be implemented in a
more » ... lines of code. First, we create a differentiable approximation of label accuracy and show that axis-aligning the bounding box around an ellipse is optimal. We then introduce Rotation Uncertainty (RU) Loss, allowing the model to adapt to the uncertainty of the labels. On five different datasets (including COCO, PascalVOC, and Transparent Object Bin Picking), this approach improves the rotational invariance of both one-stage and two-stage architectures when measured with AP, AP50, and AP75. The code is available at
arXiv:2109.13488v2 fatcat:cfroqbmdz5frzkchvyqzkdsxpa

Occluded Imaging with Time-of-Flight Sensors

Achuta Kadambi, Hang Zhao, Boxin Shi, Ramesh Raskar
2016 ACM Transactions on Graphics  
Achuta Kadambi was funded by the Charles S. Draper Doctoral Fellowship and Boxin Shi was partially supported the Singapore MOE Academic Research Fund MOE2013-T2-1-159 and the  ...  Kadambi et al. paper.  ...  Kadambi et al. Time profile imaging represents an increasingly popular research area where captured photons are parameterized by both space and time.  ... 
doi:10.1145/2836164 fatcat:7iv6z27uf5gxrnksffohtf3goy

Coded time of flight cameras

Achuta Kadambi, Refael Whyte, Ayush Bhandari, Lee Streeter, Christopher Barsi, Adrian Dorrington, Ramesh Raskar
2013 ACM Transactions on Graphics  
Glossy Object 0ns 1ns 2ns 2ns 4ns 8ns 8ns 10ns 10ns 11ns Figure 1: Using our custom time of flight camera, we are able to visualize light sweeping over the scene. In this scene, multipath effects can be seen in the glass vase. In the early time-slots, bright spots are formed from the specularities on the glass. Light then sweeps over the other objects on the scene and finally hits the back wall, where it can also be seen through the glass vase (8ns). Light leaves, first from the specularities
more » ... -10ns), then from the stuffed animals. The time slots correspond to the true geometry of the scene (light travels 1 foot in a nanosecond, times are for round-trip). Please see∼achoo/lightsweep for animated light sweep movies. Measured Amplitude Measured Range Transparent Phase Range of Foreground Range of Background Figure 2: Recovering depth of transparent objects is a hard problem in general and has yet to be solved for Time of Flight cameras. A glass unicorn is placed in a scene with a wall behind (left). A regular time of flight camera fails to resolve the correct depth of the unicorn (center-left). By using our multipath algorithm, we are able to obtain the depth of foreground (center-right) or of background (right). Abstract Time of flight cameras produce real-time range maps at a relatively low cost using continuous wave amplitude modulation and demodulation. However, they are geared to measure range (or phase) for a single reflected bounce of light and suffer from systematic errors due to multipath interference. We re-purpose the conventional time of flight device for a new goal: to recover per-pixel sparse time profiles expressed as a sequence of impulses. With this modification, we show that we can not only address multipath interference but also enable new applications such as recovering depth of near-transparent surfaces, looking through diffusers and creating time-profile movies of sweeping light. Our key idea is to formulate the forward amplitude modulated light propagation as a convolution with custom codes, record samples by introducing a simple sequence of electronic time delays, and perform sparse deconvolution to recover sequences of Diracs that correspond to multipath returns. Applications to computer vision include ranging of near-transparent objects and subsurface imaging through diffusers. Our low cost prototype may lead to new insights regarding forward and inverse problems in light transport.
doi:10.1145/2508363.2508428 fatcat:jp2cjmhjzze5rkepbvvsxtn5gm

Deep Shape from Polarization [article]

Yunhao Ba, Alex Ross Gilbert, Franklin Wang, Jinfa Yang, Rui Chen, Yiqin Wang, Lei Yan, Boxin Shi, Achuta Kadambi
2020 arXiv   pre-print
This paper makes a first attempt to bring the Shape from Polarization (SfP) problem to the realm of deep learning. The previous state-of-the-art methods for SfP have been purely physics-based. We see value in these principled models, and blend these physical models as priors into a neural network architecture. This proposed approach achieves results that exceed the previous state-of-the-art on a challenging dataset we introduce. This dataset consists of polarization images taken over a range of
more » ... object textures, paints, and lighting conditions. We report that our proposed method achieves the lowest test error on each tested condition in our dataset, showing the value of blending data-driven and physics-driven approaches.
arXiv:1903.10210v2 fatcat:u34mldx2ybb5raxjxxksrbffke

Coded aperture compressive 3-D LIDAR

Achuta Kadambi, Petros T. Boufounos
2015 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
Continuous improvement in optical sensing components, as well as recent advances in signal acquisition theory provide a great opportunity to reduce the cost and enhance the capabilities of depth sensing systems. In this paper we propose a new depth sensing architecture that exploits a fixed coded aperture to significantly reduce the number of sensors compared to conventional systems. We further develop a modeling and reconstruction framework, based on model-based compressed sensing, which
more » ... terizes a large variety of depth sensing systems. Our experiments demonstrate that it is possible to reduce the number of sensors by more than 85%, with negligible reduction on the sensing quality. Index Terms-3-D imaging, LIDAR, time of flight, compressed sensing, computational imaging.
doi:10.1109/icassp.2015.7178153 dblp:conf/icassp/KadambiB15 fatcat:vopk36khbnb73lmhio42zz34ha

3D imaging with time of flight cameras

Shahram Izadi, Ayush Bhandari, Achuta Kadambi, Ramesh Raskar
2014 ACM SIGGRAPH 2014 Courses on - SIGGRAPH '14  
Figure 1 : This course offers a self-contained tutorial to building, understanding, and exploiting Time of Flight 3D sensors. We will begin the course by introducing attendees to the landscape of 3D cameras (upper-left) that can obtain depth maps of a scene (upper-right). In the (bottom row) we show different hardware systems, which will be discussed in the course as well.
doi:10.1145/2614028.2615433 dblp:conf/siggraph/IzadiBKR14 fatcat:2dfvcb6kqjexvf4c4f7a73kyyq

3D Depth Cameras in Vision: Benefits and Limitations of the Hardware [chapter]

Achuta Kadambi, Ayush Bhandari, Ramesh Raskar
2014 Advances in Computer Vision and Pattern Recognition  
The second-generation Microsoft Kinect uses time-of-flight technology, while the first-generation Kinect uses structured light technology. This raises the question whether one of these technologies is "better" than the other. In this chapter, readers will find an overview of 3D camera technology and the artifacts that occur in depth maps. We thank the following people at the Massachusetts Institute of Technology for their contributions to the chapter:
doi:10.1007/978-3-319-08651-4_1 fatcat:gtuf2sgevzbzxf2biyzrk2zhti

Demultiplexing illumination via low cost sensing and nanosecond coding

Achuta Kadambi, Ayush Bhandari, Refael Whyte, Adrian Dorrington, Ramesh Raskar
2014 2014 IEEE International Conference on Computational Photography (ICCP)  
Several computer vision algorithms require a sequence of photographs taken in different illumination conditions, which has spurred development in the area of illumination multiplexing. Various techniques for optimizing the multiplexing process already exist, but are geared toward regular or high speed cameras. Such cameras are fast, but code on the order of milliseconds. In this paper we propose a fusion of two popular contexts, time of flight range cameras and illumination multiplexing. Time
more » ... flight cameras are a low cost, consumer-oriented technology capable of acquiring range maps at 30 frames per second. Such cameras have a natural connection to conventional illumination multiplexing strategies as both paradigms rely on the capture of multiple shots and synchronized illumination. While previous work on illumination multiplexing has exploited coding at millisecond intervals, we repurpose sensors that are ordinarily used in time of flight imaging to demultiplex via nanosecond coding strategies.
doi:10.1109/iccphot.2014.6831811 dblp:conf/iccp/KadambiBWDR14 fatcat:46voub6e65bpjbr5h2hrtwyp6m

Polarized 3D: High-Quality Depth Sensing with Polarization Cues

Achuta Kadambi, Vage Taamazyan, Boxin Shi, Ramesh Raskar
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
Coarse depth maps can be enhanced by using the shape information from polarization cues. We propose a framework to combine surface normals from polarization (hereafter polarization normals) with an aligned depth map. Polarization normals have not been used for depth enhancement before. This is because polarization normals suffer from physics-based artifacts, such as azimuthal ambiguity, refractive distortion and fronto-parallel signal degradation. We propose a framework to overcome these key
more » ... llenges, allowing the benefits of polarization to be used to enhance depth maps. Our results demonstrate improvement with respect to state-of-the-art 3D reconstruction techniques.
doi:10.1109/iccv.2015.385 dblp:conf/iccv/KadambiTSR15 fatcat:4wm3k7z5njc3jdebofk3hzevva
« Previous Showing results 1 — 15 out of 33 results