Abstract: A sensor device comprises a plurality of sensor units each of which being capable to detect an intensity of an influence on the sensor unit, and to detect as an event a positive or negative change of the intensity of the influence that is larger than a respective predetermined threshold, a filter unit configured to randomly select for readout a part of the events that were detected by the plurality of sensor units during at least one predetermined time period, a memory unit configured to store history data of detected events and/or of readout selection operations of the filter unit, and an output interface configured to receive and output the selected part of the events from the filter unit for each of the at least one predetermined time periods.

Abstract: A dynamic vision sensor (DVS) or change detection sensor reacts to changes in light intensity and in this way monitors how a scene changes. This disclosure covers both single pixel and array architectures. The DVS may contain one pixel or 2-dimensional or 1-dimensional array of pixels. The change of intensities registered by pixels are compared, and pixel addresses where the change is positive or negative are recorded and processed. Analyzing frames based on just three values for pixels, increase, decrease or unchanged, the proposed DVS can process visual information much faster than traditional computer vision systems, which correlate multi-bit color or gray level pixel values between successive frames.

Abstract: In dynamic vision sensor (DVS) or change detection sensors, the chip or sensor is configured to control or modulate the event rate. For example, this control can be used to keep the event rate close to a desired rate or within desired bounds. Adapting the configuration of the sensor to the scene by changing the ON-event and/or the OFF-event thresholds, allows having necessary amount of data, but not much more than necessary, such that the overall system gets as much information about its state as possible.

Abstract: An NMOS-only operational transconductance amplifier (OTA) replaces the PMOS transistors with switched capacitor pseudo-resistors in the pixels of an optical sensor such as a dynamic vision sensor or event-based vision sensor. Thus, if a stacked CMOS image sensor (CIS) process is employed, then the upper wafer can be kept free from N-wells, while still having the complete OTA on the upper wafer. Thus, it is possible to have only one wafer-to-wafer connection per pixel. Moreover, by operating the switched capacitor pseudo-resistors as three terminal devices, the gain can further be increased.

Abstract: A dynamic vision sensor (DVS) or change detection sensor reacts to changes in light intensity and in this way monitors how a scene changes. This disclosure covers both single pixel and array architectures. The DVS may contain one pixel or 2-dimensional or 1-dimensional array of pixels. The change of intensities registered by pixels are compared, and pixel addresses where the change is positive or negative are recorded and processed. Analyzing frames based on just three values for pixels, increase, decrease or unchanged, the proposed DVS can process visual information much faster than traditional computer vision systems, which correlate multi-bit color or gray level pixel values between successive frames.

Abstract: An event-based vision sensor system comprises an event-based pixel array including pixels that detect light, a readout circuit for reading out events associated with light received by each pixel and with the position it occupies in the array, and a memory controller that enables writing of each event, or event-related data, associated with a certain pixel to a certain address in a memory.

Abstract: In dynamic vision sensor (DVS) or change detection sensors, the chip or sensor is configured to control or modulate the event rate. For example, this control can be used to keep the event rate close to a desired rate or within desired bounds. Adapting the configuration of the sensor to the scene by changing the ON-event and/or the OFF-event thresholds, allows having necessary amount of data, but not much more than necessary, such that the overall system gets as much information about its state as possible.

Abstract: An event based pixel sensor system employing delay equalization between differently illuminated pixels.

Abstract: An event-based vision sensor system comprises an event-based pixel array including pixels that detect light, a readout circuit for reading out events associated with light received by each pixel and with the position it occupies in the array, and a memory controller that enables writing of each event, or event-related data, associated with a certain pixel to a certain address in a memory.

Abstract: A method and system for rendering an augmented reality scene on a mobile computing device tracks a real-world scene and/or a viewpoint of a user with one or more event-based vision sensors and blends an augmented reality scene displayed on the mobile computing device based on the viewpoint of the user and a scene map of the real-world scene and on the tracking of the one or more event-based vision sensors. Event-based vision sensors offer many advantages, mainly by intrinsically compressing the data stream and thus reducing the amount of data that a processing unit needs to perform. Furthermore, the event-based vision sensor pixels continuously sense the visual changes in the scene and report them with a very low latency. This makes the event-based vision sensor an ideal sensor for always-on tasks such as visual tracking and smart sensor control or data enhancement of secondary sensing modalities.

Abstract: A photoreceptor module (PR) includes a photoelectric conversion element (PD), a first amplifier circuit (110) and a second amplifier circuit (120). The first amplifier circuit (110) includes a first amplifier portion (112) and a first feedback portion (114), wherein an input of the first amplifier portion (112) and a cathode (C) of the photoelectric conversion element (PD) are electrically connected. The second amplifier circuit (120) includes a second amplifier portion (122) and a second feedback portion (124), wherein controlled path of the second feedback portion (124), a controlled path of the first feedback portion (114), and the photoelectric conversion element (PD) are electrically connected in series.

Abstract: An event based pixel sensor system employing delay equalization between differently illuminated pixels.

Abstract: A dynamic vision sensor such as an event based vision sensor employs analog to digital converters (ADC), such as ramp ADCs, that analog to digital converts the signals from photoreceptors. Only previous light values need to be stored in the pixels. This is accomplished by generating three ramps. In addition, log compression can be implemented by increasing the count linearly while increasing the reference voltage exponentially, or increasing the count logarithmically while increasing the reference voltage linearly.

Abstract: An event based pixel sensor system employs regions of interest or regions of non-interest, which are preferably dynamic, to isolate change events and reduce data from non-interesting events.

Abstract: A dynamic vision sensor such as an event based vision sensor employs analog to digital converters (ADC), such as ramp ADCs, that analog to digital converts the signals from photoreceptors. Current and previous light values are then stored and compared digitally. In addition, log compression can be implemented by increasing the count linearly while increasing the reference voltage exponentially, or increasing the count logarithmically while increasing the reference voltage linearly.

Abstract: A dynamic vision sensor (DVS) or change detection sensor reacts to changes in light intensity and in this way monitors how a scene changes. This disclosure covers both single pixel and array architectures. The DVS may contain one pixel or 2-dimensional or 1-dimensional array of pixels. The change of intensities registered by pixels are compared, and pixel addresses where the change is positive or negative are recorded and processed. Analyzing frames based on just three values for pixels, increase, decrease or unchanged, the proposed DVS can process visual information much faster than traditional computer vision systems, which correlate multi-bit color or gray level pixel values between successive frames.

Abstract: A dynamic vision sensor (DVS) or change detection sensor reacts to changes in light intensity and in this way monitors how a scene changes. This disclosure covers both single pixel and array architectures. The DVS may contain one pixel or 2-dimensional or 1-dimensional array of pixels. The change of intensities registered by pixels are compared, and pixel addresses where the change is positive or negative are recorded and processed. Analyzing frames based on just three values for pixels, increase, decrease or unchanged, the proposed DVS can process visual information much faster than traditional computer vision systems, which correlate multi-bit color or gray level pixel values between successive frames.

Abstract: A method and system for rendering an augmented reality scene on a mobile computing device tracks a real-world scene and/or a viewpoint of a user with one or more event-based vision sensors and blends an augmented reality scene displayed on the mobile computing device based on the viewpoint of the user and a scene map of the real-world scene and on the tracking of the one or more event-based vision sensors. Event-based vision sensors offer many advantages, mainly by intrinsically compressing the data stream and thus reducing the amount of data that a processing unit needs to perform. Furthermore, the event-based vision sensor pixels continuously sense the visual changes in the scene and report them with a very low latency. This makes the event-based vision sensor an ideal sensor for always-on tasks such as visual tracking and smart sensor control or data enhancement of secondary sensing modalities.

Abstract: An event-based vision sensor is fabricated using an advanced stacking technique, known as Three-Dimensional Integrated Circuit, which stacks more wafers (or dies) and interconnects them vertically. The electronic integrated circuits of the sensor are then distributed between the two or more electrically connected dies.

Abstract: A method and system for rendering an augmented reality scene on a mobile computing device tracks a real-world scene and/or a viewpoint of a user with one or more event-based vision sensors and blends an augmented reality scene displayed on the mobile computing device based on the viewpoint of the user and a scene map of the real-world scene and on the tracking of the one or more event-based vision sensors. Event-based vision sensors offer many advantages, mainly by intrinsically compressing the data stream and thus reducing the amount of data that a processing unit needs to perform. Furthermore, the event-based vision sensor pixels continuously sense the visual changes in the scene and report them with a very low latency. This makes the event-based vision sensor an ideal sensor for always-on tasks such as visual tracking and smart sensor control or data enhancement of secondary sensing modalities.