Samsung Electronics Co., Ltd.

Roy Josef Jevnisek

Ishay Goldin

Evgeny Artyomov

Michael Dinerstein

Tomer Peleg

Ran Vitek

Alexandra Dana

Netanel Stein

Matan Shoef

Yotam Perlitz

Gil Shapira

Noga Levy

Niv Zehngut

Maor Shutman

Amir Ben-dror

Methods, systems, and apparatus for combined or separate implementation of coarse-to-fine neural architecture search (NAS), two-phase block NAS, variable hardware prediction, and differential hardware design are provided and described. A variable predictor is trained, as described herein. Then, a controller or policy may be used to iteratively modify a neural network architecture along dimensions formed by neural network architecture parameters. The modification is applied to blocks (e.g., subnetworks) within the neural network architecture. In each iteration, the remainder of the neural network architecture parameters are modified and learned with a differential NAS method. The training process is performed with two-phase block NAS and incorporates a variable hardware predictor to predict power, performance, and area (PPA) parameters. The hardware parameters may be learned as well using the variable hardware predictor.

Method for co-design of hardware and neural network architectures using coarse-to-fine search, two-phased block distillation and neural hardware predictor

Embodiments of the present disclosure enable and accurate detection of facial landmarks on mobile devices in real-time. An architecture of a facial landmark detection model is provided including one or more of an attention mechanism (e.g., an attention network), a graph convolution model (e.g., a two-dimensional facial geometry graph convolution model), a multiscale coarse-to-fine mechanism, a patch-facial landmark detachment mechanism, and error estimation techniques. The attention mechanism may increase the accuracy of the facial landmark detection model by attending to meaningful patches. The graph convolution network may improve patch feature aggregation by considering the facial landmarks' geometry. The coarse-to-fine mechanism reduces a network convergence to two cycles (e.g., two facial landmark detection iterations). A patch-facial landmark detachment mechanism reduces the computation burden without significant accuracy degradation. Error estimation techniques provide accurate estimation of the regression error to the computation load and increase the accuracy of the model.

Real-time facial landmark detection

An object detection system may generate regions of interest (ROIs) from an input image that can be processed by a wide range of object detectors. According to the techniques described herein, an image is processed by a light-weight neural network (e.g., a heatmap network) that outputs object center and object scale heat-maps. The heatmaps are processed to define ROIs that are likely to include objects. Overlapping ROIs are then merged to reduce the aggregate size of the ROIs, and merged ROIs are downscaled to a reduced set of pre-defined resolutions. Fully-convolutional, high-accuracy object detectors may then operate on the downscaled ROIs to output accurate detections at a fraction of the computations by operating on a reduced image. For example, fully-convolutional, high-accuracy object detectors may operate on a subset of the entire image (e.g., cropped images based on ROIs) thus reducing computations otherwise performed over the entire image.

Region of interest selection for object detection

Growing community of inventors

Givaataim, Israel

Roy Josef Jevnisek