Methods, systems, articles of manufacture and apparatus to generate digital scenes are disclosed. An example apparatus to generate labelled models includes a map builder to generate a three-dimensional (3D) model of an input image, a grouping classifier to identify a first zone of the 3D model corresponding to a first type of grouping classification, a human model builder to generate a quantity of placeholder human models corresponding to the first zone, a coordinate engine to assign the quantity of placeholder human models to respective coordinate locations of the first zone, the respective coordinate locations assigned based on the first type of grouping classification, a model characteristics modifier to assign characteristics associated with an aspect type to respective ones of the quantity of placeholder human models, and an annotation manager to associate the assigned characteristics as label data for respective ones of the quantity of placeholder human models.

A technique for constructing a model supporting a plurality of domains is disclosed. In the technique, a plurality of teacher models, each of which is specialized for different one of the plurality of the domains, is prepared. A plurality of training data collections, each of which is collected for different one of the plurality of the domains, is obtained. A plurality of soft label sets is generated by inputting each training data in the plurality of the training data collections into corresponding one of the plurality of the teacher models. A student model is trained using the plurality of the soft label sets.

An industrial image inspection method includes: generating a test latent vector of a test image; measuring a distance between a training latent vector of a normal image and the test latent vector of the test image; and judging whether the test image is normal or defected according to the distance between the training latent vector of the normal image and the test latent vector of the test image.

In one embodiment, a method includes receiving an image on a computing device. The computing device may further execute a weakly-supervised classification algorithm to determine whether a target feature is present in the received image. As an example, the weakly-supervised classification algorithm may determine whether a building is depicted in the received image. In response to determining that a target feature is present, the method further includes using a weakly-supervised segmentation algorithm of the convoluted neural network to segment the received image for the target feature. Based on a determined footprint size of the target feature, a distribution of statistical information over the target feature in the image can be calculated.

There is provided systems and methods for training a classifier. The method comprises: obtaining a classifier for classifying data into one of a plurality of classes; retrieving training data comprising a set of observations and a set of corresponding labels, each label representing an assigned class for a corresponding observation; and applying an agent trained by a reinforcement learning system to generate labeled data from unlabeled observations and train the classifier using the training data and the labeled data according to a policy determined by the reinforcement learning system.

A computer-implemented method and systems are provided for performing machine learning for an image classification task. The method includes selecting, by a processor operatively coupled to one or more databases, a first and a second image from one or more training sets in the one or more databases. The method further includes overlaying, by the processor, the second image on the first image to form a mixed image, by averaging an intensity of each of a plurality of co-located pixel pairs in the first and the second image. The method also includes training, by the processor, a machine learning process configured for the image classification task using the mixed image to augment data used by the machine learning process for the image classification task.

Techniques and systems are described to determine personalized digital image aesthetics in a digital medium environment. In one example, a personalized offset is generated to adapt a generic model for digital image aesthetics. A generic model, once trained, is used to generate training aesthetics scores from a personal training data set that corresponds to an entity, e.g., a particular user, group of users, and so on. The image aesthetics system then generates residual scores (e.g., offsets) as a difference between the training aesthetics score and the personal aesthetics score for the personal training digital images. The image aesthetics system then employs machine learning to train a personalized model to predict the residual scores as a personalized offset using the residual scores and personal training digital images.

A system is provided for unsupervised cross-domain image generation relative to a first and second image domain that each include real images. A first generator generates synthetic images similar to real images in the second domain while including a semantic content of real images in the first domain. A second generator generates synthetic images similar to real images in the first domain while including a semantic content of real images in the second domain. A first discriminator discriminates real images in the first domain against synthetic images generated by the second generator. A second discriminator discriminates real images in the second domain against synthetic images generated by the first generator. The discriminators and generators are deep neural networks and respectively form a generative network and a discriminative network in a cyclic GAN framework configured to increase an error rate of the discriminative network to improve synthetic image quality.

A method includes: performing unsupervised pre-training of a model, the model including and a decoder including: obtaining a first image and a second image under different conditions or from different viewpoints; encoding, by the encoder, the first image into a representation of the first image and the second image into a representation of the second image; transforming the representation of the first image into a transformed representation; decoding, by the decoder, the transformed representation into a reconstructed image, where the transforming of the representation of the first image and the decoding of the transformed representation is based on the representation of the first image and the representation of the second image; and adjusting one or more parameters of at least one of the encoder and the decoder based on minimizing a loss; and fine-tuning the model, initialized with a set of task specific encoder parameters, for a geometric vision task.

A method for training an image recognition model includes: obtaining training image sets; obtaining a first predicted probability, a second predicted probability, a third predicted probability, and a fourth predicted probability based on the training image sets by using an initial image recognition model; determining a target loss function according to the first predicted probability, the second predicted probability, the third predicted probability, and the fourth predicted probability; and training the initial image recognition model based on the target loss function, to obtain an image recognition model.