Western Digital Technologies, Inc.

Sandisk Technologies Inc.

Sandisk Corporation

Rajeev Nagabhirava

Quang Le

Kuok San Ho

Xiaoyong Liu

Damian Pablo Yurzola

Zhigang Bai

Zhanjie Li

Daniele Mauri

Dhaval Parikh

Arjun Kapoor

Eran Sharon

Idan Alrod

Toshiki Hirano

Shaomin Xiong

Daniel Bai

Michael Altshuler

Madhuri Kotagiri

Itai Dror

Yong Peng

Krishnamurthy Dhakshinamurthy

Damien Kah

Yehuda Hahn

Jian Feng Chen

Thao Anh Nguyen

David Berman

Avichay Haim Hodes

Paul A Lassa

Gary Lin

Tal Heller

Matthew Davidson

Yigal Eli

Ramy Ayad

Talal Ahwal

Pritam Das

Arjun Hary

Peter Hwang

Oren Shtrasberg

Tucker Dean Berckmann

Annie Chi-San Chang

Jian Chen

Computing devices, such as mobile computing devices, have access to one or more image sensors that can capture images and video with multiple subjects. Some of these subjects may vary in priority for various tasks. It may be desired to increase or decrease the compression on each subject in order to more efficiently store the image data. Low-power, fast-response machine learning logic can be configured to allow for the generation of a plurality of inference data. Inference data can be associated with the type, motion and/or priority of the subjects as desired. This inference data can be utilized along with other subject data to generate one or more variable compression regions within the image data. The image data can be subsequently processed to compress different areas of the image based on a desired application. The variably compressed image can reduce file sizes and allow for more efficient storage and processing.

Low-power fast-response machine learning variable image compression

A digital video camera architecture for updating an object identification and tracking model deployed with the camera is disclosed. The invention comprises optics, a processor, a memory, and an artificial intelligence logic which may further comprise artificial neural networks. The architecture identifies objects according to a first confidence threshold of the model and identifies candidate objects according to the first confidence threshold and a second confidence threshold. The model may track the motion of the candidate objects within a visual field, separate the candidate objects into false positive candidate objects and false negative candidate objects according to their tracked motions, and present at least a portion of the false positive candidate objects and false negative candidate objects for further annotation.

Data collection and user feedback in edge video devices

Systems and methods for audio privacy in network video surveillance systems are described. A video camera may include an image sensor and a microphone to generate a video stream. Responsive to detecting a human speaking condition in the video stream, the audio data may be selectively modified to mask a human voice component of the audio data for storing and/or displaying the surveillance video stream.

Audio privacy protection for surveillance systems

Aspects of the present disclosure generally relate to optical devices and related methods that facilitate tilt in camera systems, such as tilt of a lens. In one example, an optical device includes a lens, an image sensor disposed below the lens, a plurality of magnets disposed about the lens, and a plurality of: (1) vertical coil structures coiled in one or more vertical planes and (2) horizontal coil structures coiled in one or more horizontal planes. When power is applied, the coil structures can generate magnetic fields that, in the presence of the magnets, cause relative movement of the coil structures and associated structures. The plurality of vertical coil structures are configured to horizontally move the lens. The plurality of horizontal coil structures are configured to tilt the lens when differing electrical power is applied to at least two of the plurality of horizontal coil structures.

Optical devices for tilt in camera systems

Systems and methods for processing surveillance video streams using image classification and object detection are described. Video data from a video image sensor may be processed using an image classifier to determine whether an object type is present in a video frame. If the object type is present, the video frame and/or subsequent video frames may be processed using an object detector to provide additional object data, such as position information, for use in other video surveillance processes. In some examples, an event message may be generated and sent to a video surveillance application in response to selective object detection.

Processing of surveillance video streams using image classification and object detection

Computing devices, such as mobile computing devices, have access to one or more image sensors that can capture images with multiple subjects. Some of these subjects may be known to the user capturing an image with the image sensor. The user may prefer to have the captured image data be optimized around the known subjects. Low-power, fast-response machine learning logic can be configured to allow for the generation of a plurality of inference data. This inference data can be utilized along with other sensor data, such as a motion sensor, for the generation of one or more image sensor configuration changes that may be implemented to optimize the subsequent capture of image data. This cycle of image data analysis, image sensor optimization, and subsequent capture can continue multiple times until a threshold of optimization or time is met. The captured image data optimized around the known subjects is then stored.

Low-power, fast-response machine learning autofocus enhancements

The present disclosure generally relates to a method and device for reduce the amount of energy used to revert data within a storage device to a consistent state after a power loss event. Once power is lost from a host device, and prior to complete shutdown, a log is created within non-volatile memory (NVM) of the storage device. The log contains logical block addresses (LBAs) corresponding to data that experienced a write failure. At the next power-on event, the log is checked to see if any LBAs are present. If LBAs are present in the log, then the data that experienced a write failure can be properly written to the storage device.

Delayed write failure logging

The present disclosure generally relates to sensor device, such as a magnetic sensor bridge, that utilizes a dual free layer (DFL) structure. The device includes a plurality of resistors that each includes the same DFL structure. Adjacent the DFL structure is a magnetic structure that can include a permanent magnet, an antiferromagnetic (AFM) layer having a synthetic AFM (SAF) structure thereon, a permanent magnetic having a SAF structure thereon, or an AFM layer having a ferromagnetic layer thereon. The DFL structures are aligned with different layers of the magnetic structures to differentiate the resistors. The different alignment and/or different magnetic structures result in a decrease in production time due to reduced complexity and, thus, reduces costs.

Magnetic sensor bridge using dual free layer

Aspects of the present disclosure relate to optical devices and related methods that facilitate independent control of movement of lenses and image sensors in camera systems. In one example, an image sensor is movable independently of and relative to a lens, and the lens is movable independently of the image sensor. In one example, an optical device includes a lens, and an image sensor disposed below the lens. The image sensor is movable relative to the lens. The optical device includes a plurality of magnets disposed about the lens, a plurality of vertical coil structures coiled in one or more vertical planes, and one or more horizontal coil structures coiled in one or more horizontal planes. The plurality of vertical coil structures are configured to, when powered, move the image sensor relative to the lens. The one or more horizontal coil structures are configured to, when powered, move the lens.

Optical devices for independent movement control of lenses and image sensors in camera systems

In a flash memory, redundant columns are used alternatively as replacement columns for replacing bad columns or to provide additional redundancy for ECC encoding. Locations of bad columns are indicated to a soft-input ECC decoder so that data bits from bad columns are treated as having a lower reliability than data bits from other columns.

Bad column handling in flash memory

A memory device and techniques for its operation are presented. After operating on power received from a host, the memory device determines that it is no longer receiving host power and, in response, activates a power source on the memory device itself. Using this reserve power, the memory device can then perform data management operations. The techniques can also be applied to a digital appliance having a non-volatile memory. The memory device or digital appliance can prioritize its memory management operation during the host/user operating window based on the ability to perform these operations outside of the host/user operating window. Additionally, in a data write operations, where the memory device receives data from a host, stores the data in volatile memory, and then writes the data into the non-volatile memory, the memory device sends the host an acknowledgment of the data having been written into the non-volatile memory after it has been store in the volatile memory, but before the write into the non-volatile memory is complete.

Leveraging portable system power to enhance memory management and enable application level features

Growing community of inventors

Santa Clara, CA, United States of America

Rajeev Nagabhirava