Qualcomm Incorporated

Seungwoo Yoo

Hee-Seok Lee

Duck Hoon Kim

Heesoo Myeong

Kang Min Kim

Hyun-Mook Cho

Young-Ki Baik

Jihoon Kim

Te-Won Lee

Kyu Woong Hwang

Sungrack Yun

Duck-hoon Kim

Sungwoong Kim

Jeong-Kyun Lee

Seok-Soo Hong

Jimin Jeon

Certain aspects of the present disclosure provide techniques for lane marker detection. A set of feature tensors is generated by processing an input image using a convolutional neural network. A set of localizations is generated by processing the set of feature tensors using a localization network, a set of horizontal positions is generated by processing the set of feature tensors using row-wise regression, and a set of end positions is generated by processing the set of feature tensors using y-end regression. A set of lane marker positions is determined based on the set of localizations, the set of horizontal positions, and the set of end positions.

Lane marker recognition

Certain aspects of the present disclosure provide a method for lane marker detection, including: receiving an input image; providing the input image to a lane marker detection model; processing the input image with a shared lane marker portion of the lane marker detection model; processing output of the shared lane marker portion of the lane marker detection model with a plurality of lane marker-specific representation layers of the lane marker detection model to generate a plurality of lane marker representations; and outputting a plurality of lane markers based on the plurality of lane marker representations.

Lane marker detection

Disclosed are techniques for performing lane instance recognition. Lane instances are difficult to recognize since they are long and elongated, and they also look different from view to view. An approach is proposed in which local mask segmentation lane estimation and global control points lane estimation are combined.

Lane marker detection and lane instance recognition

Certain aspects of the present disclosure are directed to methods and apparatus for deep learning in an artificial neural network. One example method generally includes receiving input data at an input to a layer of the neural network; replicating a group of neural processing units in the layer to form a superset of neural processing units, the superset comprising n instances of the group of neural processing units; processing the input data using the superset to generate output data for the layer; and determining an uncertainty of the output data. Processing the input data includes performing a dropout function by zeroing out one or more weights of a set of weights for each of the n instances of the superset of neural processing units and convolving, for each of the n instances in parallel, the input data with one or more non-zeroed out weights of the set of weights.

Efficient dropout inference for bayesian deep learning

A method performed by an apparatus is described. The method includes receiving remote data. The method also includes adapting model processing based on the remote data. The method further includes determining a driving decision based on the adapted model processing. In some examples, adapting model processing may include selecting a model, adjusting model scheduling, and/or adjusting a frame computation frequency.

Systems and methods for adaptive model processing

Disclosed are techniques for estimating multiple lane boundaries through simultaneous detection of lane markers (LM) and raised pavement markers (RPM). Conventional techniques for lane marker detection (LMD) comprises extracting and clustering line segments from a camera image, fitting the clustered lines to a geometric model, and selecting the multiple lanes using heuristic approaches. Unfortunately, in the conventional technique, an error from each step is sequentially propagated to the next step. Also there are no techniques to estimate both LMs and RPMs. A technique to simultaneously detect in real time both LM and RPMs is proposed. This enables optimal estimation of multiple lane boundaries.

Real-time simultaneous detection of lane marker and raised pavement marker for optimal estimation of multiple lane boundaries

An electronic device is described. The electronic device includes a memory and a processor in communication with the memory. The processor is configured to receive an image. The processor is also configured to extract a lane response map from the image. The processor is further configured to estimate one or more lane markers based on the lane response map.

Systems and methods for lane-marker detection

A method, performed by an electronic device, for linking a thumbnail of an image and at least one web page is disclosed. In this method, the image including at least one text region may be accessed in a storage unit. At least one text region may be detected in the image and at least one character string in the at least one text region may be recognized. Further, the method may include selecting the at least one web page from the plurality of web pages and linking the thumbnail of the image and the at least one web page.

Linking thumbnail of image to web page

A method, which is performed by an electronic device, for resizing an image having text is disclosed. The method may include determining layout information of at least one text region in the image. The layout information may include at least one of a number, a size, a location, a shape, or a text density of the at least one text region in the image. The method may also select a seam carving operation, a cropping operation, or a scaling operation for the image based on the layout information, a size of the image, and a target image size. The selected operation may be performed to resize the image to the target image size based at least on one of the layout information, the size of the image, or the target image size. The resized image may include the at least one text region.

Text-based image resizing

A method, performed by an electronic device, for verifying a user to allow access to the electronic device is disclosed. In this method, sensor data may be received from a plurality of sensors including at least an image sensor and a sound sensor. Context information of the electronic device may be determined based on the sensor data and at least one verification unit may be selected from a plurality of verification units based on the context information. Based on the sensor data from at least one of the image sensor or the sound sensor, the at least one selected verification unit may calculate at least one verification value. The method may determine whether to allow the user to access the electronic device based on the at least one verification value and the context information.

Context-based access verification

A method, which is performed by an electronic device, of automatically activating a flash for an image sensor of the electronic device is disclosed. The method may include receiving a first image including at least one text region and determining feature data characterizing the at least one text region in the first image. The method may also identify at least one candidate specular reflection region in the first image. Based on the feature data and the at least one candidate specular reflection region, the flash may be activated. Upon activating the flash, a second image including the at least one text region may be captured.

Activating flash for capturing images with text

According to an aspect of the present disclosure, a method for adjusting a view mirror in a vehicle is disclosed. The method includes obtaining a first angle of the view mirror, capturing an image of a head of a driver, determining, from the captured image, a viewing distance and a perpendicular distance between a location in the head and the view mirror, calculating a view angle based on the viewing distance and the perpendicular distance, wherein the view angle is an angle between a direction orthogonal to the view mirror and a view direction associated with the location in the head and the view mirror, determining a second angle of the view mirror based on the first angle and the view angle, and adjusting the view mirror from the first angle to the second angle.

Growing community of inventors

Yongin-si, South Korea

Seungwoo Yoo