Ford Global Technolgoies, LLC

Jack Cooper Logistics, LLC

Mohit Prabhushankar

Andrea Amico

Ghassan AlRegib

Mostafa Parchami

Enrique Corona

Dogancan Temel

Armin Parchami

Gukyeong Kwon

Ryan Benkert

Neural networks and learning algorithms can use a variance of gradients to provide a heuristic understanding of the model. The variance of gradients can be used in active learning techniques to train a neural network. Techniques include receiving a dataset with a vector. The dataset can be annotated and a loss calculated. The loss value can be used to update the neural network through backpropagation. An updated dataset can be used to calculate additional losses. The loss values can be added to a pool of gradients. A variance of gradients can be calculated from the pool of gradient vectors. The variance of gradients can be used to update a neural network.

Variance of gradient based active learning framework for training perception algorithms

A deep neural network (DNN) can be trained based on a first training dataset that includes first images including annotated first objects. The DNN can be tested based on the first training dataset to determine first object predictions including first uncertainties. The DNN can be tested by inputting a second training dataset and outputting first object predictions including second uncertainties, wherein the second training dataset includes second images including unannotated second objects. A subset of images included in the second training dataset can be selected based on the second uncertainties, The second objects in the selected subset of images included in the second training dataset can be annotated. The DNN can be trained based on the selected subset of images included in the second training dataset including the annotated second objects.

Neural network training

A portable computing device equipped with at least one image capture device and/or a light source captures an image (or a video) of a portion of a surface of interest having the damage that is exposed to a light from the light source. The portable computing device converts the image to an output image that highlights the damage. If the damage is a dent, the image is converted to a false color image using a saliency algorithm. If the damage is a scratch, the image is converted to a colorspace stretched image using color stretching algorithms. The size of the damage is determined by capturing an image of a ruler placed adjacent to the damage and the portion of surface of interest having the damage. The ruler is then removed from the image. The resulting image is converted to the output image. The ruler is added to the output image.

System, method, and apparatus for detection of damages on surfaces

A portable computing device equipped with an image capture device captures an image of a vehicle dashboard of a vehicle. Then, the portable computing device identifies the location of one or more components of the vehicle dashboard in the captured image. Based on the location of the one or more components, the portable computing device segments the captured image to obtain an image of each of the one or more components. Further, the portable computing device processes the images of the one or more components using one or more machine learning models to determine a reading associated with each of the one or more components. An accuracy of the readings is verified and responsively, the portable computing device inputs the readings in respective data fields of an electronic form. The readings associated with the one or more components of the vehicle dashboard represent data associated with the vehicle.

Growing community of inventors

Atlanta, GA, United States of America

Mohit Prabhushankar