Objectron: A Large Scale Dataset of Object-Centric Videos in the Wild with Pose Annotations
CVPR-2021, abs: https://arxiv.org/abs/2012.09988
pdf: https://arxiv.org/pdf/2012.09988
project page: https://objectron.dev
3D object detection has recently become popular due to many applications in robotics, augmented reality, autonomy, and image retrieval. We introduce the Objectron dataset to advance the state of the art in 3D object detection and foster new research and applications, such as 3D object tracking, view synthesis, and improved 3D shape representation. The dataset contains object-centric short videos with pose annotations for nine categories and includes 4 million annotated images in 14,819 annotated videos. We also propose a new evaluation metric, 3D Intersection over Union, for 3D object detection. We demonstrate the usefulness of our dataset in 3D object detection tasks by providing baseline models trained on this dataset. Our dataset and evaluation source code are available online at http://www.objectron.dev
Instant 3D Object Tracking with Applications in Augmented Reality
Forth Workshop on Computer Vision for AR/VR in Conference on Computer Vision and Pattern Recognition, 2020. [pdf]
Tracking object poses in 3D is a crucial building block for Augmented Reality applications. We propose an instant
motion tracking system that tracks an object’s pose in space (represented by its 3D bounding box) in real-time on mobile devices. Our system does not require any prior sensory calibration or initialization to function. We employ a deep neural network to detect objects and estimate their initial 3D pose. Then the estimated pose is tracked using a robust planar tracker. Our tracker is capable of performing relative-scale 9-DoF tracking in real-time on mobile devices. By combining use of CPU and GPU efficiently, we achieve 26-FPS+ performance on mobile devices
MobilePose: Real-Time Pose Estimation for Unseen Objects with Weak Shape Supervision
In this paper, we address the problem of detecting unseen objects from RGB images and estimating their poses in 3D. We propose two mobile friendly networks: MobilePose-Base and MobilePose-Shape. The former is used when there is only pose supervision, and the latter is for the case when shape supervision is available, even a weak one. We revisit shape features used in previous methods, including segmentation and coordinate map. We explain when and why pixel-level shape supervision can improve pose estimation. Consequently, we add shape prediction as an intermediate layer in the MobilePose-Shape, and let the network learn pose from shape. Our models are trained on mixed real and synthetic data, with weak and noisy shape supervision. They are ultra lightweight that can run in real-time on modern mobile devices (e.g. 36 FPS on Galaxy S20). Comparing with previous single-shot solutions, our method has higher accuracy, while using a significantly smaller model (2~3% in model size or number of parameters).
Instant Motion Tracking and Its Applications to Augmented Reality
Third Workshop on Computer Vision for AR/VR in Conference on Computer Vision and Pattern Recognition, 2019. [pdf]
Augmented Reality (AR) brings immersive experiences to users. With recent advances in computer vision and mo- bile computing, AR has scaled across platforms, and has increased adoption in major products. One of the key chal- lenges in enabling AR features is proper anchoring of the virtual content to the real world, a process referred to as tracking. In this paper, we present a system for motion tracking, which is capable of robustly tracking planar tar- gets and performing relative-scale 6DoF tracking without calibration. Our system runs in real-time on mobile phones and has been deployed in multiple major products on hun- dreds of millions of devices.
Validation and Optimization of Analog Circuits using Randomized Search Algorithms
My Ph.D. thesis, published by University of Illinois at Urbana-Champaign. [pdf][Duplex]
The thesis page and the abstract
Duplex: Simultaneous Parameter-Performance Exploration for Optimizing Analog Circuits
[paper] – Proceedings of 2016 International Conference On Computer Aided Design (ICCAD), 2016.
We present Duplex random tree search, an algorithm to optimize performance metrics of analog and mixed signal circuits. Duplex determines the optimal design, the Pareto set and the sensitivity of circuit’s performance metrics to its parameters. We demonstrate that Duplex is 5× faster than the state-of-the-art and finds the global optimum for a design whose previously published result was a local optimum. We show our algorithm’s scalability by optimizing a system-level post-layout charged-pump PLL circuit.
A Random Tree Search Algorithm for Nash Equilibrium in Capacitated Selfish Replication Games
[paper][source] – Proceedings of IEEE 55th Conference on Decision and Control (CDC), 2016
In this paper, we consider a resource allocation game with binary preferences and limited capacities over large scale networks and propose a novel randomized algorithm for searching its pure-strategy Nash equilibrium points. It is known that such games always admit a pure-strategy Nash equilibrium and benefit from having a low price of anarchy. However, the best known theoretical results only provide a quasi-polynomial constant approximation algorithm of the equilibrium points over general networks. Here, we search the state space of the resource allocation game for its equilibrium points. We use a random tree based search method to minimize a proper score function and direct the search toward the pure-strategy Nash equilibrium points of the system. We demonstrate efficiency of our algorithm through some empirical results.
A Novel Test Compression Algorithm for Analog Circuits to Decrease Production Costs
[pdf] – Published in Elsevier VLSI integration journal, 2016.
Minimizing the manufacturing test time for ICs is one of the main keys to reducing the product cost. We introduce a methodology for automated test compression for electrical stress testing of analog and mixed signal circuits. This methodology optimally extracts only portions of a functional test that electrically stress the nets and devices of an analog circuit. We model test compression as a problem of optimizing functional of the transient response. We present a random tree based approach to find the minimum for these computationally hard integrals, which corresponds to the optimally compressed analog test. We demonstrate with an op-amp, VCO, and CMOS inverter that the method consistently reduces the length of each test by an average of 93%. Our technology can compress tests in the presence of process variation and utilize parallel processing to speed up the compression algorithm.
Every Test Makes a Difference: Compressing Analog Tests to Decrease Production Costs
[pdf] –Proceedings of 21st Asia and South Pacific Design Automation Conference (ASP-DAC 2016), 2016.
Minimizing the manufacturing test time for ICs is one of the main keys to reducing the product cost.
We introduce a methodology for automated test compression during electrical stress testing of analog and mixed signal circuits. This methodology optimally extracts only portions of a functional test that electrically stress the nets and devices of an analog circuit. We model test compression as a problem of optimizing functionals of the transient response. We present a random tree based approach to find optimal solutions for these computationally hard integrals. We demonstrate with an op-amp, VCO and CMOS inverter that the method consistently reduces the length of each test by 93%.