CVDec 16, 2016

Fast, Dense Feature SDM on an iPhone

arXiv:1612.05332v1
Originality Incremental advance
AI Analysis

This enables real-time facial or object tracking on mobile devices, but is incremental as it builds on existing SDM methods.

The paper tackled the problem of running dense Supervised Descent Method (SDM) on mobile devices by proposing a Sparse Compositional Regression (SCR) framework and Binary Approximated SIFT (BASIFT) features, achieving over 90 FPS on an iPhone 7 with similar accuracy to SDM.

In this paper, we present our method for enabling dense SDM to run at over 90 FPS on a mobile device. Our contributions are two-fold. Drawing inspiration from the FFT, we propose a Sparse Compositional Regression (SCR) framework, which enables a significant speed up over classical dense regressors. Second, we propose a binary approximation to SIFT features. Binary Approximated SIFT (BASIFT) features, which are a computationally efficient approximation to SIFT, a commonly used feature with SDM. We demonstrate the performance of our algorithm on an iPhone 7, and show that we achieve similar accuracy to SDM.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes