Abstract: A system and method are disclosed for computer vision in-store planogram analysis. The system includes a retail entity comprising a product display area defined by a planogram, and a planogram analyzer comprising a server which trains a machine learning model to generate recommendations to optimize one or more planogram metrics using historical data and computer vision data. The system generates, by the trained machine learning model, recommendations to optimize the one or more planogram metrics, and alters the planogram based on the one or more recommendations. The system further uses the trained machine learning model with a probabilistic optimization technique approximating a global optimum to generate the one or more recommendations to optimize the one or more planogram metrics. The computer vision includes one or more of eyeball movement, eyeball engagement and eyeball lingering time.

Abstract: A system and method of automatic product attribute recognition receive training images having bounding boxes associated with one or more products in the training images, receive attribute values for each of the one or more products in the training images, and train a first convolutional neural network (CNN) model to generate bounding boxes for and identify each of the one or more products with the training images until the accuracy of the first CNN model is above a first predetermined threshold. The system and method further train a second CNN model for each of the products associated with the cropped images until the second CNN generates attribute values for the one or more attributes with an accuracy above a second predetermined threshold, and automatically recognize the one or more attributes for a new product image by presenting the product image to the first and second CNN models.

Abstract: A system and method of automatic product attribute recognition receive training images having bounding boxes associated with one or more products in the training images, receive attribute values for each of the one or more products in the training images, and train a first convolutional neural network (CNN) model to generate bounding boxes for and identify each of the one or more products with the training images until the accuracy of the first CNN model is above a first predetermined threshold. The system and method further train a second CNN model for each of the products associated with the cropped images until the second CNN generates attribute values for the one or more attributes with an accuracy above a second predetermined threshold, and automatically recognize the one or more attributes for a new product image by presenting the product image to the first and second CNN models.

Abstract: A system and method of automatic product attribute recognition receive training images having bounding boxes associated with one or more products in the training images, receive attribute values for each of the one or more products in the training images, and train a first convolutional neural network (CNN) model to generate bounding boxes for and identify each of the one or more products with the training images until the accuracy of the first CNN model is above a first predetermined threshold. The system and method further train a second CNN model for each of the products associated with the cropped images until the second CNN generates attribute values for the one or more attributes with an accuracy above a second predetermined threshold, and automatically recognize the one or more attributes for a new product image by presenting the product image to the first and second CNN models.

Abstract: A system and method of automatic product attribute recognition receive training images having bounding boxes associated with one or more products in the training images, receive attribute values for each of the one or more products in the training images, and train a first convolutional neural network (CNN) model to generate bounding boxes for and identify each of the one or more products with the training images until the accuracy of the first CNN model is above a first predetermined threshold. The system and method further train a second CNN model for each of the products associated with the cropped images until the second CNN generates attribute values for the one or more attributes with an accuracy above a second predetermined threshold, and automatically recognize the one or more attributes for a new product image by presenting the product image to the first and second CNN models.

Abstract: A system and method are disclosed for computer vision in-store planogram analysis. The system includes a retail entity comprising a product display area defined by a planogram, and a planogram analyzer comprising a server which trains a machine learning model to generate recommendations to optimize one or more planogram metrics using historical data and computer vision data. The system generates, by the trained machine learning model, recommendations to optimize the one or more planogram metrics, and alters the planogram based on the one or more recommendations. The system further uses the trained machine learning model with a probabilistic optimization technique approximating a global optimum to generate the one or more recommendations to optimize the one or more planogram metrics. The computer vision includes one or more of eyeball movement, eyeball engagement and eyeball lingering time.

Abstract: A system and method of automatic product attribute recognition receive training images having bounding boxes associated with one or more products in the training images, receive attribute values for each of the one or more products in the training images, and train a first convolutional neural network (CNN) model to generate bounding boxes for and identify each of the one or more products with the training images until the accuracy of the first CNN model is above a first predetermined threshold. The system and method further train a second CNN model for each of the products associated with the cropped images until the second CNN generates attribute values for the one or more attributes with an accuracy above a second predetermined threshold, and automatically recognize the one or more attributes for a new product image by presenting the product image to the first and second CNN models.

Abstract: This disclosure relates generally to ROI marking, and more particularly to system and method for marking ROI in a media stream using touchless hand gesture interface such as headmount devices. In one embodiment, the method includes recognizing a pointing object representative of a gesture in frames of the media stream while capturing the media stream. The media stream comprises a scene captured from a first person view (FPV) of a user. Locus of the pointing object is detected in subsequent frames subsequent of the media stream to select a ROI in the media stream. The locus of the pointing object configures a bounding box around the ROI. The ROI is tracked in frames of the media stream occurring subsequent to the subsequent frames in the media stream. The bounding box is updated around the ROI based on the tracking, wherein the updated bounding box encloses the ROI.

Abstract: This disclosure relates generally to hand-gesture recognition, and more particularly to system and method for detecting interaction of 3D dynamic hand gestures with frugal AR devices. In one embodiment, a method for hand-gesture recognition includes receiving frames of a media stream of a scene captured from a FPV of a user using RGB sensor communicably coupled to a wearable AR device. The media stream includes RGB image data associated with the frames of the scene. The scene comprises a dynamic hand gesture performed by the user. Temporal information associated with the dynamic hand gesture is estimated from the RGB image data by using a deep learning model. The estimated temporal information is associated with hand poses of the user and comprises key-points identified on user's hand in the frames. Based on said temporal information, the dynamic hand gesture is classified into predefined gesture classes by using multi-layered LSTM classification network.

Abstract: This disclosure relates generally to hand-gesture recognition, and more particularly to system and method for detecting interaction of 3D dynamic hand gestures with frugal AR devices. In one embodiment, a method for hand-gesture recognition includes receiving frames of a media stream of a scene captured from a FPV of a user using RGB sensor communicably coupled to a wearable AR device. The media stream includes RGB image data associated with the frames of the scene. The scene comprises a dynamic hand gesture performed by the user. Temporal information associated with the dynamic hand gesture is estimated from the RGB image data by using a deep learning model. The estimated temporal information is associated with hand poses of the user and comprises key-points identified on user's hand in the frames. Based on said temporal information, the dynamic hand gesture is classified into predefined gesture classes by using multi-layered LSTM classification network.

Abstract: This disclosure relates generally to ROI marking, and more particularly to system and method for marking ROI in a media stream using touchless hand gesture interface such as headmount devices. In one embodiment, the method includes recognizing a pointing object representative of a gesture in frames of the media stream while capturing the media stream. The media stream comprises a scene captured from a first person view (FPV) of a user. Locus of the pointing object is detected in subsequent frames subsequent of the media stream to select a ROI in the media stream. The locus of the pointing object configures a bounding box around the ROI. The ROI is tracked in frames of the media stream occurring subsequent to the subsequent frames in the media stream. The bounding box is updated around the ROI based on the tracking, wherein the updated bounding box encloses the ROI.